From 7a742e7a65222463d9fbc3743970b78ffc525624 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Wed, 5 Nov 2025 09:17:11 -0600 Subject: [PATCH 01/19] Ref Arch for Unifying Access Across for SAP Business Data Cloud using SAP Cloud Identity Services --- .../drawio/cloud-identity-services-bdc.drawio | 403 ++++++++++++++++++ .../6-cloud-identity-services-bdc/readme.md | 245 +++++++++++ 2 files changed, 648 insertions(+) create mode 100644 docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio create mode 100644 docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md diff --git a/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio new file mode 100644 index 0000000000..0bdf4cf919 --- /dev/null +++ b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio @@ -0,0 +1,403 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + \ No newline at end of file diff --git a/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md new file mode 100644 index 0000000000..0b8ed85be7 --- /dev/null +++ b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md @@ -0,0 +1,245 @@ +--- +id: id-ra0013-6 +slug: /ref-arch/f5b6b597a6/6 +sidebar_position: 3 +title: Unifying Access Across SAP BDC with SAP Cloud Identity Services +description: >- + Unifying Access Across for SAP Business Data Cloud using SAP Cloud Identity Services: IAS for SSO (SAML/OIDC) and IPS for SCIM provisioning. Includes scenarios with/without Enterprise IdP, lifecycle, authorization mapping, and operations. +keywords: + - sap business data cloud + - sap databricks + - sap cloud identity services + - ias + - ips + - sso + - saml + - oidc + - scim 2.0 + - unity catalog + - zero trust +sidebar_label: Unifying Access Across SAP BDC with SAP Cloud Identity Services +tags: + - data + - aws + - azure + - gcp +hide_table_of_contents: false +hide_title: false +toc_min_heading_level: 2 +toc_max_heading_level: 4 +contributors: + - guilherme-segantini + - jmsrpp + - anbazhagan-uma +last_update: + author: guilherme-segantini + date: 2025-11-04 +--- + +This reference architecture proposes a solution to unify access across the client's landscape with SAP Business Data Cloud (BDC), including SAP Analytics Cloud, SAP Datasphere, and SAP Databricks. By leveraging **SAP Cloud Identity Services (CIS)**—including IAS for SSO and IPS for provisioning—you can establish a single source of truth for user identities, groups of access, streamline access control, and enhance security. + +## Why Unified Identity: Key Outcomes + +- **Improved Governance and Consistency**: A unified identity model ensures that access policies are consistently applied across all systems, reducing the risk of security breaches and simplifying audits. +- **Increased Business Agility**: By automating the user lifecycle, you can onboard new users and grant them access to the resources they need in a timely manner, accelerating time-to-value. +- **Reduced Operational Costs**: Centralizing identity management reduces the administrative overhead associated with managing multiple systems, freeing up IT resources to focus on more strategic initiatives. +- **Enhanced Security Posture**: A unified identity model provides a single source of truth for user identities, making it easier to detect and respond to security threats. +- **Future-Proof Architecture**: By leveraging open standards such as SAML, OIDC, and SCIM, you can ensure that your identity management solution is able to adapt to changing business needs and new technologies. + +### User Journey: Intelligent App Developer (Alex) + +To ground the architecture in a real use case, we introduce the persona "Alex," an intelligent app developer. We use this persona to clarify why unified identity is needed and to illustrate the end-to-end journey across SAP BDC, from first assignment to governed data access. + +:::note[Introducing Alex] +Alex builds data-driven apps that combine using multiple tools, including SAP Analytics Cloud (SAC), governed data products from BDC, and advanced AI notebooks (SAP Databricks). As a member of the `intelligent-app-developers` enterprise group, he expects **consistent privileges** across every tool—meaning the same group yields the same effective rights in SAC roles, Datasphere spaces/roles, SAP BTP role collections, Databricks workspace entitlements, and Unity Catalog data access. He also expects one login (SSO via IAS), immediate workspace access (pre-provisioned via IPS/SCIM), and zero ticket waiting time (IGA-approved automated JML), with MFA and conditional access enforced uniformly and auditable access with centralized logs. +::: + +## Architecture and Design Principles for Unified Identity + +Our approach uses SAP Cloud Identity Services (IAS and IPS) as a central broker to create a consistent identity and authorization model across all platforms. This is guided by the following core principles: + +- **Central Governance**: Establish a single source of truth for identities to enforce the principle of least privilege. +- **Separation of Duties**: Create clear boundaries between platform administration and data governance roles. +- **Zero Trust**: Treat every access request as untrusted. Enforce MFA and conditional access at the network edge. +- **Audit & Remediate**: Log all access, regularly audit for unauthorized changes, and automate remediation. +- **Open Standards**: Leverage SAML, OIDC, and SCIM to ensure a flexible and future-proof architecture. + +## Architecture + +The architecture positions SAP Cloud Identity Services as the central hub for identity management. It unifies access by connecting to an authoritative source of truth, which can be an Enterprise IdP or a core business system like **SAP SuccessFactors** (for employee lifecycle) or **SAP S/4HANA** (identity data or business roles). Any such system that can be integrated with SAP Identity Provisioning Service (IPS) and serve as the authoritative source for user lifecycle events. CIS then provides single sign-on (SSO) and user provisioning to SAP BDC. + +![drawio](drawio/cloud-identity-services-bdc.drawio) + +### Key Components + +- **SAP Cloud Identity Services (CIS)**: The core of this architecture, providing a unified identity and access management solution. +- **SAP Identity Authentication Service (IAS)**: Acts as the central SSO broker, handling authentication requests and federating to other Enterprise IdPs as needed. +- **SAP Identity Provisioning Service (IPS)**: Automates user lifecycle management, ensuring that identities and access groups are consistent across all connected systems. +- **SAP BTP Subaccount**: The Cloud Foundry environment hosting the native BDC applications. It serves as the central provisioning target, where IPS syncs users and groups. These identities are then available to all applications within the subaccount, like SAP Analytics Cloud, SAP Datasphere, and the BDC Cockpit. Authorization is managed here via Role Collections. +- **SAP Databricks**: The data and AI platform, which is provisioned to separately. It has its own account-level SCIM directory for user provisioning and uses workspace entitlements and Unity Catalog for authorization. + +## Providing unified access for SAP BDC + +:::note[Our persona: IPS pre-provisions Alex's access across SAP BDC (SCIM)] +IPS automatically syncs Alex's IAS group memberships to each target system. When Alex signs in via IAS, his token maps directly to the correct BTP role collections, Databricks workspace entitlements, and Unity Catalog grants. +::: + +For Alex, seamless access across SAP Business Data Cloud means one secure sign-on through IAS and pre-provisioned, group-based authorization via IPS applied consistently in SAP Analytics Cloud, SAP Datasphere, and SAP Databricks. The following flow diagrams use SAP Databricks as an example, but the same identity patterns and governance model apply across BDC applications. + +### Access Provisioning Flow + +1. **Group Definition**: A group, **`intelligent-app-developers`**, is defined. This can be in the Enterprise Directory and replicated via IPS, or created directly in the IAS user store. +2. **User Assignment**: Alex is made a member of the `intelligent-app-developers` group. +3. **Provisioning to Databricks**: An IPS job is configured with IAS as the source and SAP Databricks as the target. It reads the `intelligent-app-developers` group and its members. +4. **SCIM Synchronization**: IPS provisions the group and its members to the SAP Databricks account-level SCIM directory. + +```mermaid +sequenceDiagram + participant Dir as Enterprise Directory + participant IAS as SAP IAS + participant IPS as SAP IPS + participant DBX as SAP Databricks + + alt Group Master is Enterprise Directory + Dir->>IPS: Read group 'intelligent-app-developers' + IPS->>IAS: Create/update group + else Group Master is IAS + IAS-->>IAS: Admin creates 'intelligent-app-developers' + end + + IAS-->>IPS: Read group 'intelligent-app-developers' for provisioning + IPS->>DBX: Provision "intelligent-app-developers" group via SCIM +``` + +:::note[Key Takeaways] +- **Proactive Entitlements**: Groups are provisioned to Databricks *before* users log in for the first time. +- **Automated Access**: New members of the `intelligent-app-developers` group automatically get access without manual intervention. +- **Flexible Governance**: Group management can be centralized in your Enterprise Directory or managed directly within SAP Cloud Identity Services. For SAC, Datasphere, and the BDC Cockpit, admins typically map SAP BTP role collections to IAS groups once per application/role collection. This mapping only needs updates when new role collections or groups are introduced, names change, or governance requirements evolve. +::: + +### Authentication & Authorization Flow + +:::note[User Journey: Single sign-on across SAP BDC (IAS SSO)] +Alex signs in once with IAS (SAML/OIDC). IAS can broker to the enterprise IdP as needed, so one login grants seamless access across BDC apps. +::: + +1. **Single Sign-On**: Alex launches the SAP Databricks or any BDC application (e.g., SAP Analytics Cloud, SAP Datasphere), which redirects to IAS for authentication. +2. **Authentication**: Alex signs in. This may be federated to an Enterprise IdP or handled directly by IAS. +3. **Token Issuance**: IAS issues a SAML or OIDC token containing Alex's group memberships, including **`intelligent-app-developers`**. +4. **Accessing Databricks**: The token is presented to SAP Databricks. +5. **Authorization**: Databricks recognizes the `intelligent-app-developers` group (which has been pre-provisioned) and grants Alex the appropriate permissions within the workspace and Unity Catalog. + +```mermaid +sequenceDiagram + participant Dir as Opt. Enterprise Directory + participant IAS as SAP IAS + participant IPS as IPS (IAS→Databricks) + participant DBX as SAP Databricks + participant Alex as Alex + + alt Group Source is Enterprise Directory + Dir->>IAS: Sync group `intelligent-app-developers` (via IPS) + else Group Source is IAS + IAS-->>IAS: Admin creates `intelligent-app-developers` + end + + Alex-->>IAS: Is made member of group + + IPS->>DBX: Provision "intelligent-app-developers" group via SCIM + + Alex->>DBX: Attempts to log in + DBX->>IAS: SAML/OIDC Request + IAS-->>IAS: Authenticates Alex + IAS->>DBX: SAML/OIDC Token with group membership + DBX-->>Alex: Access granted based on group +``` + +### Databricks Identity Types + +In a CIS-centered architecture, Databricks identities are integrated via IAS for SSO (SAML/OIDC) and IPS for SCIM provisioning. The same three types apply, but their lifecycle and authorization are governed by IAS/IPS: + +- **Users (human identities)**: + - **Source of truth**: Enterprise Directory (e.g., SuccessFactors/AD) synchronized into IAS, or IAS-native users. + - **Authentication**: Single sign-on via IAS (SAML or OIDC), with federation to the Enterprise IdP as needed. Enforce conditional access and MFA in IAS. + - **Provisioning**: IPS pushes user objects to the Databricks account-level SCIM directory. Ensure email/username normalization and disable local invites to keep Databricks aligned with IAS. + - **Authorization**: Users inherit access through pre-provisioned IAS groups that IPS syncs to Databricks. Effective rights materialize as workspace entitlements and Unity Catalog privileges granted to those groups. + +- **Service principals (non-human identities)**: + - **Purpose**: CI/CD pipelines, scheduled jobs, automation tooling without interactive login. + - **Provisioning**: Prefer SCIM-based creation via IPS where supported; alternatively create in Databricks and bind access through IAS-managed groups that IPS syncs. + - **Credentials**: Favor OAuth client credentials or cloud-native identities over PATs. Store and rotate secrets via platform vaults; scope to least privilege in Unity Catalog. + - **Governance**: Treat as privileged identities with dedicated approval flows and separation from human-user groups. + +- **Groups (authorization control plane)**: + - **Mastering**: Define groups in IAS (or Enterprise Directory) and synchronize to Databricks via IPS SCIM. + - **Mapping**: Align IAS groups to Databricks workspace entitlements (e.g., “Can Use”, “Workspace Admin”) and to Unity Catalog grants on catalogs/schemas/tables. + - **Design**: Use clear naming conventions that tie business roles to technical privileges; keep groups atomic; avoid direct user grants in Databricks—favor group-based RBAC (and ABAC with tags, where applicable). + +## Audit & Monitoring + +SAP Cloud Identity Services (IAS for SSO, IPS for provisioning) serve as the central identity and lifecycle control plane, anchoring audit and monitoring across the BDC landscape. + +- **Unified identity context** — Understand who accessed and under which policies; resource-level details (e.g., which story, space, data product or table) come from SAC, Datasphere, and Databricks logs correlated with CIS. +- **Consistent policy enforcement** — One set of SSO and conditional access controls (IAS) to monitor for compliance and zero trust. +- **Lifecycle traceability** — Provisioned users and groups (IPS) reflect business intent, making access changes explainable and auditable. +- **Cross-platform governance** — A common identity model aligns application and data platform events to business roles. +- **Faster detection and response** — Correlated identity and authorization signals shorten time to detect, investigate, and prove compliance. + +## Deployment Scenarios + +:::note[User Journey: Alex] +Select the rollout option that keeps him productive, whether greenfield or enterprise-brokered. +::: + +The following scenarios are designed to help you choose the best deployment strategy for your organization, based on your size and existing identity landscape. + +### Greenfield Scenarios + +**For Small to Medium-Sized Enterprises (SMEs):** + +* **Scenario A: SAP Cloud Identity Services as the Primary Enterprise IdP** + * **Description**: In this scenario, SAP Cloud Identity Services (CIS) is used as the central Enterprise IdP for both SSO and user provisioning. This is the simplest and most straightforward approach for organizations that do not have an existing Enterprise IdP. + * **Recommendation**: This is the recommended approach for SMEs that are starting from scratch with their identity management solution. + +**For Large Enterprises:** + +* **Scenario B: Enterprise IdP with SAP Cloud Identity Services as a Broker** + * **Description**: In this scenario, the organization's existing Enterprise IdP is used as the authoritative source for user identities, and CIS is used as a broker to provide SSO and user provisioning to SAP BDC and Databricks. + * **Recommendation**: This is the recommended approach for large enterprises that have an existing Enterprise IdP and want to leverage their existing investment. + +### Brownfield Scenarios + +**For Small to Medium-Sized Enterprises (SMEs):** + +* **Scenario C: Hybrid Approach** + * **Description**: In this scenario, a hybrid approach is used, where some applications are integrated directly with the Enterprise IdP, while others are integrated with CIS. This approach can be used to migrate to a fully unified identity management solution over time. + * **Recommendation**: This approach is recommended for SMEs that have an existing Enterprise IdP but are not yet ready to migrate all of their applications to a unified identity management solution. + +**For Large Enterprises:** + +* **Scenario D: Phased Rollout** + * **Description**: In this scenario, a phased rollout is used, where the unified identity management solution is implemented for a small number of applications and then gradually rolled out to the rest of the organization. + * **Recommendation**: This approach is recommended for large enterprises that want to minimize the risk of disruption to their business operations. + +## Best-Practice Checklist + +- **Least Privilege** – Grant only required role-collections and Unity Catalog privileges. +- **MFA Everywhere** – Enforce multi-factor authentication via IAS conditional access. +- **Centralized Logging** – Stream all IAS, IPS, BTP, and BDC audit logs into SIEM and review routinely. +- **Access Certification** – Conduct regular reviews with business data owners. + +## Conclusion +:::note[User Journey: Alex] +Unified identity gives instant access, consistent privileges, and lets him focus on building apps—not chasing permissions. +::: +This reference architecture has provided a comprehensive overview of how to implement a unified identity and access management solution for SAP Business Data Cloud and SAP Databricks. By leveraging SAP Cloud Identity Services, you can establish a single source of truth for user identities, streamline access control, and enhance security. + +The deployment scenarios and recommendations provided in this document are designed to help you choose the best approach for your organization, based on your size and existing identity landscape. By following the guidance in this document, you can implement a robust and scalable identity management solution that will help you deliver analytics and AI confidently at scale. + +## References +- [SAP Cloud Identity Services](https://discovery-center.cloud.sap/serviceCatalog/cloud-identity-services?region=all) +- [SAP Identity Authentication Service (IAS) documentation](https://help.sap.com/docs/identity-authentication) +- [SAP Identity Provisioning Service (IPS) documentation](https://help.sap.com/docs/identity-provisioning) +- [SAP Business Data Cloud overview](https://www.sap.com/products/technology-platform/business-data-cloud.html) +- [SAP BTP Authorization and Trust Management](https://help.sap.com/docs/btp/sap-business-technology-platform/sap-authorization-and-trust-management-service-in-cloud-foundry-environment?q=Authorization+Service+) +- [Create or Modify a Bundled SAP Cloud Identity Services Tenant](https://help.sap.com/docs/business-data-cloud/administering-sap-business-data-cloud/create-or-modify-bundled-sap-cloud-identity-services-tenant) From dfc65ac3507413f1624b06f99c0c8b9b5bbd85c7 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Wed, 5 Nov 2025 12:02:40 -0600 Subject: [PATCH 02/19] Removing draft references not done yet --- .../6-sap-databricks-sap-aicore/readme.md | 44 ------------------- .../RA0013/7-sap-bdc-security/readme.md | 44 ------------------- 2 files changed, 88 deletions(-) delete mode 100644 docs/ref-arch/RA0013/6-sap-databricks-sap-aicore/readme.md delete mode 100644 docs/ref-arch/RA0013/7-sap-bdc-security/readme.md diff --git a/docs/ref-arch/RA0013/6-sap-databricks-sap-aicore/readme.md b/docs/ref-arch/RA0013/6-sap-databricks-sap-aicore/readme.md deleted file mode 100644 index 2b660fe158..0000000000 --- a/docs/ref-arch/RA0013/6-sap-databricks-sap-aicore/readme.md +++ /dev/null @@ -1,44 +0,0 @@ ---- -id: id-ra0013-6 -slug: /ref-arch/f5b6b597a6/6 -sidebar_position: 6 -sidebar_custom_props: - category_index: [] -title: SAP Databricks and SAP AI Core Integration -description: >- - SAP and Databricks have partnered to integrate SAP data with Databricks AI and - analytics platform, allowing businesses to leverage SAP data for AI and - machine learning applications. This partnership simplifies data access and - eliminates the need for complex ETL processes, enabling real-time analytics - and AI-driven decision-making. -keywords: - - sap - - bdc - - business - - data - - cloud - - databricks -sidebar_label: SAP Databricks and SAP AI Core Integration -image: img/ac-soc-med.png -tags: - - data - - aws - - azure - - gcp -hide_table_of_contents: false -hide_title: false -toc_min_heading_level: 2 -toc_max_heading_level: 4 -draft: true -unlisted: false -contributors: - - anbazhagan-uma - -discussion: -last_update: - author: anbazhagan-uma - date: 2025-08-10 ---- - -# SAP AI Core and Databricks Integration in SAP Business Data Cloud - This content is currently under discussion for external release. diff --git a/docs/ref-arch/RA0013/7-sap-bdc-security/readme.md b/docs/ref-arch/RA0013/7-sap-bdc-security/readme.md deleted file mode 100644 index eca238d8da..0000000000 --- a/docs/ref-arch/RA0013/7-sap-bdc-security/readme.md +++ /dev/null @@ -1,44 +0,0 @@ ---- -id: id-ra0013-7 -slug: /ref-arch/f5b6b597a6/7 -sidebar_position: 7 -sidebar_custom_props: - category_index: [] -title: SAP BDC Security and Compliance -description: >- - SAP and Databricks have partnered to integrate SAP data with Databricks AI and - analytics platform, allowing businesses to leverage SAP data for AI and - machine learning applications. This partnership simplifies data access and - eliminates the need for complex ETL processes, enabling real-time analytics - and AI-driven decision-making. -keywords: - - sap - - bdc - - business - - data - - cloud - - databricks -sidebar_label: SAP BDC Security and Compliance -image: img/ac-soc-med.png -tags: - - data - - aws - - azure - - gcp -hide_table_of_contents: false -hide_title: false -toc_min_heading_level: 2 -toc_max_heading_level: 4 -draft: true -unlisted: false -contributors: - - anbazhagan-uma - -discussion: -last_update: - author: anbazhagan-uma - date: 2025-08-10 ---- - -# SAP BDC Security and Compliance - This content is currently WIP for external release. From 4c2855c4519a272fc04c6721549c1e8f1f54c725 Mon Sep 17 00:00:00 2001 From: Uma Anbazhagan Date: Thu, 6 Nov 2025 14:06:25 +0530 Subject: [PATCH 03/19] minor change - gui repo --- .../drawio/cloud-identity-services-bdc.drawio | 806 +++++++++--------- .../6-cloud-identity-services-bdc/readme.md | 18 +- 2 files changed, 413 insertions(+), 411 deletions(-) diff --git a/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio index 0bdf4cf919..29b981c78a 100644 --- a/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio +++ b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/drawio/cloud-identity-services-bdc.drawio @@ -1,403 +1,403 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - \ No newline at end of file + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md index 0b8ed85be7..7a154e8595 100644 --- a/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md +++ b/docs/ref-arch/RA0013/6-cloud-identity-services-bdc/readme.md @@ -1,7 +1,7 @@ --- id: id-ra0013-6 slug: /ref-arch/f5b6b597a6/6 -sidebar_position: 3 +sidebar_position: 6 title: Unifying Access Across SAP BDC with SAP Cloud Identity Services description: >- Unifying Access Across for SAP Business Data Cloud using SAP Cloud Identity Services: IAS for SSO (SAML/OIDC) and IPS for SCIM provisioning. Includes scenarios with/without Enterprise IdP, lifecycle, authorization mapping, and operations. @@ -19,14 +19,16 @@ keywords: - zero trust sidebar_label: Unifying Access Across SAP BDC with SAP Cloud Identity Services tags: - - data - - aws - - azure - - gcp + - data + - sap bdc + - security + - sap-databricks hide_table_of_contents: false hide_title: false toc_min_heading_level: 2 toc_max_heading_level: 4 +draft: false +unlisted: false contributors: - guilherme-segantini - jmsrpp @@ -54,7 +56,7 @@ To ground the architecture in a real use case, we introduce the persona "Alex," Alex builds data-driven apps that combine using multiple tools, including SAP Analytics Cloud (SAC), governed data products from BDC, and advanced AI notebooks (SAP Databricks). As a member of the `intelligent-app-developers` enterprise group, he expects **consistent privileges** across every tool—meaning the same group yields the same effective rights in SAC roles, Datasphere spaces/roles, SAP BTP role collections, Databricks workspace entitlements, and Unity Catalog data access. He also expects one login (SSO via IAS), immediate workspace access (pre-provisioned via IPS/SCIM), and zero ticket waiting time (IGA-approved automated JML), with MFA and conditional access enforced uniformly and auditable access with centralized logs. ::: -## Architecture and Design Principles for Unified Identity +## Design Principles for Unified Identity for SAP BDC Our approach uses SAP Cloud Identity Services (IAS and IPS) as a central broker to create a consistent identity and authorization model across all platforms. This is guided by the following core principles: @@ -66,7 +68,7 @@ Our approach uses SAP Cloud Identity Services (IAS and IPS) as a central broker ## Architecture -The architecture positions SAP Cloud Identity Services as the central hub for identity management. It unifies access by connecting to an authoritative source of truth, which can be an Enterprise IdP or a core business system like **SAP SuccessFactors** (for employee lifecycle) or **SAP S/4HANA** (identity data or business roles). Any such system that can be integrated with SAP Identity Provisioning Service (IPS) and serve as the authoritative source for user lifecycle events. CIS then provides single sign-on (SSO) and user provisioning to SAP BDC. +The architecture positions SAP Cloud Identity Services as the central hub for identity management. It unifies access by connecting to an authoritative source of truth, which can be an Enterprise IdP or a core business system like **SAP SuccessFactors** (for employee lifecycle) or **SAP S/4HANA** (identity data or business roles). Any such system that can be integrated with SAP Identity Provisioning Service (IPS) and serve as the authoritative source for user lifecycle events. SAP CIS then provides single sign-on (SSO) and user provisioning to SAP BDC. ![drawio](drawio/cloud-identity-services-bdc.drawio) @@ -225,7 +227,7 @@ The following scenarios are designed to help you choose the best deployment stra - **Least Privilege** – Grant only required role-collections and Unity Catalog privileges. - **MFA Everywhere** – Enforce multi-factor authentication via IAS conditional access. -- **Centralized Logging** – Stream all IAS, IPS, BTP, and BDC audit logs into SIEM and review routinely. +- **Centralized Logging** – Stream all SAP IAS, SAP IPS, SAP BTP, and SAP BDC audit logs into SIEM and review routinely. - **Access Certification** – Conduct regular reviews with business data owners. ## Conclusion From 4e7353bf04d15499c8fca8ef397e81c2a9b6a28c Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Fri, 5 Dec 2025 16:42:28 -0600 Subject: [PATCH 04/19] Revised reference architecture for BDC/AI Core --- docs/learning jorney/Unit4.md | 363 ++++++++++++++++++ .../RA0013/7-bdc-powered-by-ai-core/readme.md | 202 ++++++++++ 2 files changed, 565 insertions(+) create mode 100644 docs/learning jorney/Unit4.md create mode 100644 docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md diff --git a/docs/learning jorney/Unit4.md b/docs/learning jorney/Unit4.md new file mode 100644 index 0000000000..86d64a52fb --- /dev/null +++ b/docs/learning jorney/Unit4.md @@ -0,0 +1,363 @@ +# Designing Hybrid and Customized Architectures for Enterprise Architects +## Strategic Foundations Course + +## Course Overview + +This entry-level strategic course equips enterprise architects with fundamental knowledge to design hybrid architectures that integrate SAP and non-SAP systems, leverage AI capabilities, and ensure data reliability. Participants will learn to apply SAP Architecture Center reference architectures as practical blueprints for customer implementations, focusing on strategic decision-making and architectural assessment. + +**Target Audience:** Enterprise Architects (entry level) +**Duration:** 90-120 minutes +**Focus:** Strategic fundamentals, not deep technical implementation + +--- + +## Course Structure + +### Lesson 1: Hybrid Architecture Strategy (30 minutes) + +#### Learning Objectives: +By the end of this lesson, participants will be able to: +- Explain the strategic value of hybrid architectures for enterprise transformation +- Identify when to use cloud, on-premise, and multi-cloud patterns +- Understand SAP Business Data Cloud as a unified data platform +- Leverage SAP Architecture Center reference architectures for decision-making + +#### 1.1 Why Hybrid Architectures Matter (10 minutes) + +**Strategic Drivers:** +- **Business Agility:** Rapid deployment of new capabilities without replacing legacy systems +- **Data Sovereignty:** Compliance with regional regulations (GDPR, data residency) +- **Cost Optimization:** Balance cloud economics with existing infrastructure investments +- **Risk Mitigation:** Gradual migration reduces disruption + +**Key Decision Points:** +- Which workloads belong in cloud vs. on-premise? +- How to maintain data consistency across environments? +- What integration patterns minimize complexity? + +**SAP Architecture Center Value:** +- Pre-validated blueprints reduce architecture risk +- Standardized patterns accelerate time-to-value +- Alignment with SAP's strategic roadmap + +**Reference:** [RA0008: Edge Integration Cell](https://architecture.learning.sap.com/docs/ref-arch/263f576c90) + +#### 1.2 SAP Business Data Cloud - Strategic Overview (15 minutes) + +**What is SAP BDC?** +A unified platform that brings together SAP and non-SAP data for analytics and AI: +- **SAP Datasphere:** Connect and harmonize data from any source +- **SAP Analytics Cloud:** Business intelligence and planning +- **SAP Databricks:** Advanced AI/ML capabilities + +**Strategic Capabilities:** +- **Data Products:** Reusable, governed data assets +- **Federation vs. Replication:** Choose based on latency needs and cost +- **Zero-Copy Architecture:** Access data without moving it (via Delta Share) + +**When to Use BDC:** +- ✓ Need unified view across SAP and non-SAP systems +- ✓ Building AI/ML solutions requiring high-quality data +- ✓ Modernizing analytics infrastructure +- ✗ Simple reporting from single SAP system (consider native tools first) + +**Reference:** [RA0013: SAP Business Data Cloud](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6) + +#### 1.3 Integration and Security Fundamentals (5 minutes) + +**Integration Strategy:** +- **API-First:** SAP Integration Suite for orchestration +- **Event-Driven:** Real-time data flows for critical processes +- **Hybrid Execution:** Edge Integration Cell for on-premise connectivity + +**Security Essentials:** +- Zero-trust principles (verify every access) +- Identity and access management (covered in Lesson 3) +- Secure connectivity (Cloud Connector, Private Link) + +**Reference:** [RA0001: Event-Driven Architecture](https://architecture.learning.sap.com/docs/ref-arch/fbdc46aaae) + +--- + +### Lesson 2: AI Agent Strategy and Interoperability (30 minutes) + +#### Learning Objectives: +By the end of this lesson, participants will be able to: +- Understand when and why to use AI agents in enterprise architectures +- Distinguish between SAP's two interoperability standards: MCP and A2A +- Apply Agent2Agent (A2A) protocol for cross-platform agent collaboration +- Make strategic decisions about agent deployment in hybrid landscapes + +#### 2.1 AI Agents - Strategic Perspective (10 minutes) + +**What are AI Agents?** +Autonomous systems that can reason, plan, and execute multi-step tasks with minimal human intervention. + +**When to Consider AI Agents:** +- ✓ Complex workflows spanning multiple systems +- ✓ Tasks requiring contextual decision-making +- ✓ Processes benefiting from natural language interaction +- ✗ Simple automation (use traditional workflow tools) +- ✗ Single-step operations (use standard APIs) + +**Two Development Paths:** +1. **Content-Based (Joule Studio):** Low-code, business-user friendly, rapid deployment +2. **Code-Based (LangGraph, AutoGen):** Full control, complex custom logic, developer-led + +**Strategic Question:** Which approach fits your organization's skills and timeline? + +**Reference:** [RA0005: AI Agents & Agent Builder](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d/5) + +#### 2.2 Interoperability Strategy: MCP vs. A2A (15 minutes) + +**Why Interoperability Matters:** +In hybrid landscapes, SAP Joule agents need to collaborate with third-party agents (Microsoft Copilot, Google Vertex AI, AWS Bedrock). SAP's strategy uses two standards with distinct purposes: + +**1. Model Context Protocol (MCP) - Internal Use Only** +- **Purpose:** Connect AI agents to SAP data and tools +- **Scope:** Internal to SAP systems +- **Key Point:** Not for customer architecture design - SAP manages MCP internally + +**2. Agent-to-Agent (A2A) - External Collaboration** +- **Purpose:** Enable agents from different vendors to work together +- **Scope:** Cross-platform, vendor-neutral standard +- **Key Point:** This is what enterprise architects should design with + +**Strategic Architecture Principle:** +> Use A2A as the standard pattern for agent collaboration in hybrid architectures. Ensure SAP remains the governed data hub, with external agents interacting via A2A protocol only. + +**Why A2A Matters:** +- **Security:** Agents collaborate without exposing internal data access +- **Governance:** Centralized control through Agent & Tool Gateway +- **Flexibility:** Plug-and-play ecosystem of specialized agents +- **Standardization:** Vendor-neutral, future-proof approach + +**Architect's Decision Framework:** +- Third-party agent integration? → Use A2A protocol +- Data access from agents? → Route through SAP BDC with governance +- Custom agent development? → Consider Joule Studio first, code-based if needed + +**Reference:** [RA0005: Agent Interoperability](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d/8) + +#### 2.3 Data Access Strategy for Agents (5 minutes) + +**Two Primary Data Access Patterns:** + +1. **Structured Data (SQL-based)** + - Agents query SAP Datasphere using natural language + - Federated access across SAP and non-SAP sources + - Role-based security enforced + +2. **Unstructured Data (RAG - Retrieval Augmented Generation)** + - Vector search in SAP HANA Cloud for document similarity + - Grounding AI responses in enterprise knowledge + - Reduces hallucination, improves accuracy + +**Governance Essentials:** +- Data lineage tracking +- Access audit trails +- Compliance (GDPR, industry regulations) + +**Reference:** [RA0005: Retrieval Augmented Generation](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d/3) + +--- + +### Lesson 3: Assessing SAP Business Data Cloud Architecture and Engineering for Increased AI Reliability (30 minutes) + +#### Learning Objectives: +By the end of this lesson, participants will be able to: +- Assess SAP Business Data Cloud architectures using a structured framework +- Understand identity and access management strategy for BDC +- Apply key reliability principles for AI systems +- Make informed architecture decisions based on trade-off analysis + +#### 3.1 BDC Architecture Assessment Framework (15 minutes) + +**Strategic Assessment Dimensions:** + +When evaluating a BDC architecture, consider these five critical areas: + +1. **Scalability** + - Can it handle 10x data growth? + - Concurrent user capacity adequate? + - Strategy: Federation (flexible) vs. Replication (performance) + +2. **Security** + - Zero-trust implementation (verify every access) + - Data encryption (at rest and in transit) + - Identity and access management maturity + +3. **Reliability** + - Uptime targets (99.9% = 8.7 hours downtime/year) + - Data freshness requirements (real-time vs. batch) + - Disaster recovery plan (RPO/RTO defined?) + +4. **Cost Optimization** + - Federation saves storage, but may increase compute + - Replication costs more storage, but faster queries + - Data lifecycle policies to archive cold data + +5. **AI/ML Readiness** + - Data quality sufficient for model training? + - Feature engineering capabilities available? + - Production deployment infrastructure ready? + +**Key Trade-off Decisions:** + +| Decision | Option A | Option B | Choose A When... | Choose B When... | +|----------|----------|----------|------------------|------------------| +| Data Access | Federation | Replication | Data freshness critical, storage expensive | Query performance critical, budget allows | +| Processing | Real-time | Batch | Latency <1 sec required | Cost optimization priority, can tolerate delay | +| Data Products | SAP-managed | Customer-managed | Standard use cases, faster deployment | Custom logic, full control needed | + +**Reference:** [RA0013: SAP Business Data Cloud](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6) + +#### 3.2 Identity and Access Management Strategy (10 minutes) + +**SAP Cloud Identity Services (CIS):** + +**Core Components:** +- **IAS (Identity Authentication):** Single sign-on across SAP and BDC +- **IPS (Identity Provisioning):** Automated user lifecycle management + +**Strategic Deployment Patterns:** + +1. **Greenfield (New Implementation)** + - Use IAS as primary identity provider + - Simpler setup, faster deployment + - Best for: SMEs, new SAP landscape + +2. **Brownfield (Existing Enterprise IdP)** + - Federate enterprise IdP (Azure AD, Okta) with IAS + - IAS acts as broker to SAP systems + - Best for: Large enterprises, existing IAM infrastructure + +**Zero-Trust Principles:** +- Multi-factor authentication (MFA) mandatory +- Least privilege access (only what's needed) +- Continuous validation (not "trust once") +- Centralized audit logging + +**Why This Matters for AI:** +- AI agents need controlled access to data +- Granular permissions prevent data leakage +- Audit trails for compliance and debugging + +**Reference:** [RA0019: Identity and Access Management](https://architecture.learning.sap.com/docs/ref-arch/20c6b29b1e) + +#### 3.3 AI Reliability Essentials (5 minutes) + +**Key Reliability Principles:** + +1. **Data Quality** + - Automated validation at ingestion + - Monitor for drift (data distribution changes) + - Schema evolution management + +2. **Model Lifecycle** + - Separate environments (dev, staging, prod) + - Gradual rollout (canary deployments) + - Automated retraining when accuracy drops + +3. **Monitoring** + - Track: Latency, accuracy, throughput + - Alert on: Performance degradation, errors + - Review: Regular postmortem analysis + +4. **Governance** + - Model approval workflows + - Bias detection and fairness metrics + - Compliance (GDPR, regulatory requirements) + +**Critical Questions for Architects:** +- What's the acceptable downtime for AI services? +- How quickly must we detect model degradation? +- Who approves model deployments to production? +- How do we ensure AI decisions are explainable? + +**Reference:** [RA0005: Generative AI on SAP BTP](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d) + +--- + +## Key SAP Architecture Center Reference Architectures + +### Essential References: +- **[RA0001: Event-Driven Architecture (EDA)](https://architecture.learning.sap.com/docs/ref-arch/fbdc46aaae)** +- **[RA0005: Generative AI on SAP BTP](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d)** +- **[RA0008: Edge Integration Cell](https://architecture.learning.sap.com/docs/ref-arch/263f576c90)** +- **[RA0013: SAP Business Data Cloud](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6)** (Primary Reference) +- **[RA0019: Identity and Access Management](https://architecture.learning.sap.com/docs/ref-arch/20c6b29b1e)** +- **[RA0024: Integrating and Extending Joule](https://architecture.learning.sap.com/docs/ref-arch/06ff6062dc)** + +### Supporting References: +- **[RA0004: Modernizing SAP BW with SAP BDC](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6/4)** +- **[RA0012: HANA Cloud Medallion Architecture](https://architecture.learning.sap.com/docs/ref-arch/d9b25daf96)** +- **[RA0026: Embodied AI Agents](https://architecture.learning.sap.com/docs/ref-arch/083f2d968e)** + +--- + +## Course Takeaways + +### What You've Learned: + +1. **Hybrid Architecture Design** + - Integration patterns for cloud, on-premise, and multi-cloud + - Data federation vs. replication strategies + - Secure connectivity and zero-trust principles + +2. **AI Agent Development & Interoperability** + - Content-based vs. code-based development approaches + - Agent2Agent protocol for cross-platform collaboration + - Data access strategies for structured and unstructured data + +3. **SAP BDC for AI Reliability** + - Unified data platform architecture + - Identity and access management with Cloud Identity Services + - AI governance and compliance frameworks + +### How to Apply SAP Architecture Center: + +- **Start with Reference Architectures:** Use proven blueprints as starting points +- **Adapt to Context:** Customize patterns for your specific requirements +- **Follow Best Practices:** Leverage SAP-recommended approaches +- **Stay Current:** Monitor updates to reference architectures +- **Engage Community:** Contribute and learn from peer implementations + +--- + +## Next Steps + +### Immediate Actions: +1. Explore SAP Architecture Center for your specific use cases +2. Identify hybrid integration requirements in your landscape +3. Assess AI agent opportunities for automation +4. Review identity and access management maturity + +### Advanced Learning: +- Deep-dive workshops on specific reference architectures +- Hands-on labs with SAP BTP services +- Architecture review sessions with SAP experts +- Implementation guidance and best practices + +### Resources: +- SAP Architecture Center: Comprehensive reference architectures +- SAP Discovery Center: Guided missions and tutorials +- SAP Help Documentation: Technical specifications +- SAP Community: Blogs, forums, and peer insights + +--- + +## Additional Resources + +### Documentation: +- [SAP Architecture Center](https://www.sap.com/architecture-center) +- [SAP Business Data Cloud](https://www.sap.com/products/technology-platform/business-data-cloud.html) +- [SAP Cloud Identity Services](https://discovery-center.cloud.sap/serviceCatalog/cloud-identity-services) +- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) + +### Sample Implementations: +- GitHub: SAP-samples repositories +- SAP Discovery Center: Mission-based learning paths +- SAP Community: Blog posts and tutorials + +--- diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md new file mode 100644 index 0000000000..a9840a4dda --- /dev/null +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -0,0 +1,202 @@ +--- +id: id-ra0013-7 +slug: /ref-arch/f5b6b597a6/7 +sidebar_position: 7 +title: SAP Business Data Cloud powered by SAP AI Core +description: >- + Strategic value and architectural patterns for integrating SAP Business Data Cloud with SAP AI Core and Generative AI Hub. Covers AI-Enhanced Data Products, model training in Databricks and serving in AI Core, batch and real-time consumption patterns, predictive insights, and autonomous process optimization with AI agents. +keywords: + - sap business data cloud + - sap ai core + - generative ai hub + - tabular ai + - sap databricks + - data products + - model lifecycle + - enterprise ai + - reference architecture +sidebar_label: SAP Business Data Cloud powered by SAP AI Core +tags: + - ai + - data + - sap-ai-core + - genai + - databricks + - mlops + - architecture + - data-products +hide_table_of_contents: false +hide_title: false +toc_min_heading_level: 2 +toc_max_heading_level: 4 +draft: false +unlisted: false +contributors: + - SeeObjective + - guilherme-segantini + - jmsrpp + - anbazhagan-uma +last_update: + author: guilherme-segantini + date: 2025-12-05 +--- + +Enterprises possess a wealth of invaluable business data within their SAP systems. However, activating this data for modern Artificial Intelligence is often a complex, disconnected, and risky endeavor. To stay competitive, organizations need a strategy to transform this data into reliable, governed, and actionable AI-driven insights that are deeply integrated with core business processes. + +This reference architecture presents a cohesive vision for combining **SAP Business Data Cloud** with **SAP AI Foundation** (including SAP AI Core and the Generative AI Hub). The core architectural concept is the creation of **AI-Enhanced Data Products**—intelligent, context-aware, and dynamic assets that deliver trusted predictive insights and drive business automation at scale. This integrated approach enables the development of AI solutions that are **reliable, relevant, and responsible**, accelerating time-to-value and embedding intelligence directly into the enterprise. + +## The Architectural Blueprint + +The architectural vision is centered on a powerful synergy: SAP Business Data Cloud provides the governed, business-ready data foundation, while SAP AI Foundation offers a comprehensive portfolio of enterprise-grade AI capabilities. This avoids a "one-size-fits-all" approach, recognizing that different business problems require different tools. The value lies in the comphreensive AI capabilities that allows data scientists and developers to move from data preparation to model deployment and inference. + +This blueprint is built on the principle of using each component for its primary strength: + +- **SAP Business Data Cloud (with SAP Databricks):** A comprehensive **enterprise data platform** that serves as the unified foundation for AI-driven business intelligence. It provides end-to-end capabilities for **discovering, connecting, preparing, exploring, curating, and governing** both SAP and non-SAP data sources through a semantically rich, business-context-aware layer. + +- **SAP AI Foundation (SAP AI Core & Generative AI Hub):** A comprehensive **enterprise AI platform** that offers a complete set of capabilities for managing the full lifecycle of machine learning and generative AI models. It supports scalable training, deployment, and monitoring, along with workflow orchestration, model versioning, and secure integration with SAP applications. SAP AI Core enables customers to use several models or allow customers to bring and operationalize their own AI models. These models can be accessed and governed through the Generative AI Hub, making them reusable services across the enterprise. + +## Design Principles and Unique Strengths of SAP AI Core + +* **Rich Portfolio of GenAI Models:** Delivered via the Generative AI Hub and governed as reusable services. This provides a broader selection of foundation models tailored for SAP-centric workloads than is available on many general-purpose platforms. + +* **SAP-First Governance:** Model training, deployment, monitoring, and auditability are all aligned with SAP's identity, policy, and compliance expectations, ensuring enterprise-grade security. + +* **Operational Alignment:** Features such as deployment targets, access controls, observability, and cost management are specifically designed to integrate smoothly with existing SAP applications and business processes. + +* **Complementary by Design:** SAP AI Core is built to work alongside, not replace, existing data platforms. It complements lakehouse analytics and data science environments by operationalizing AI services *inside* the SAP landscape, which accelerates value realization. + +## Key Architectural Patterns + +To compete today, enterprises must activate their most valuable asset — their core business data in SAP — for modern AI. However, this is often a complex, disconnected, and risky endeavor. + +The strategy we're outlining provides a clear, governed path to solve this. It's built on the powerful synergy between SAP Business Data Cloud for our data foundation and SAP AI Foundation for our AI capabilities. + +### Pattern 1: Train in Databricks, Serve in AI Core (The "Foundational Pattern") + +**What:** A data scientist performs exploratory data science and model training in SAP Databricks, directly against governed, business-ready data products shared from SAP BDC Cockpit. During development, they engineer features within their notebooks—simple transformations stay with the model as preprocessing steps, while reusable features can be promoted back to BDC data products for broader consumption. The goal is to produce a high-quality, production-ready model. The key is a clean separation of roles: the data scientist creates the model or utilizes an existing model and releases the model artifacts to the unity catalog in Databricks, while a client-managed MLOps pipeline pulls these artifacts and handles validation and deployment to AI Core using SAP AI Core SDK. + +**Why:** This pattern provides separation of concerns—data scientists use a pro-code development environment to run experiments in Databricks while AI Core handles the production model serving. + +### Pattern 2: The "Batch Consumption" Pattern (for Data Enrichment) + +**What:** Building on Pattern 1's deployed model, this pattern enriches data products through high-throughput batch processing using scheduled or event-driven triggers (e.g., hourly ticket categorization, weekly sales forecasts). The model reads from a source dataset and writes predictions back—either as new columns in the existing dataset or as a separate data product in BDC. + +**Trigger Options:** The automated pipeline can decide when batch processes run, providing full control and flexibility for integration: +* **Recurring:** Schedule periodic batch jobs to continuously process new or updated records as they arrive. BDC's Delta Lake foundation enables efficient incremental processing through Change Data Feed (CDF), ensuring only new data is processed rather than rescanning entire datasets. +* **One-Time:** Trigger bulk processing for historical data or initial data product creation across entire datasets. +* **Event-Driven:** Build logic to automatically trigger re-scoring in response to events, such as the deployment of a new model (via Pattern 1), ensuring insights reflect the latest AI logic. + +**Workflow Implementation:** Use **SAP AI Core workflow executions** for batch processing—distinct from real-time endpoints and optimized for high-throughput operations. Your **containerized workflow** reads from configured object stores (S3, Azure Blob), processes data, and writes predictions back. Schedule jobs with **cron in SAP AI Launchpad** for simple recurring tasks, or use the **SAP AI Core SDK** for client-managed programmatic control and integration with existing pipelines. + +**Why:** This architecture provides a robust, governable, and efficient way to handle all non-real-time AI processing. It uses the *same, single, governed model* from AI Core for all batch scenarios, ensuring that whether you are processing a small batch of new records or re-scoring an entire table, you are using the same "reliable" and "responsible" AI logic. It separates high-throughput batch workloads from low-latency online serving, ensuring the right resources are used for the right job, promoting both performance and cost-efficiency. + +### Pattern 3: The "Real-Time Consumption" Pattern (for Embedded Intelligence) + +**What:** Building on Pattern 1's deployed model, this pattern embeds AI into live business processes for immediate predictions. Applications make low-latency API calls to the **deployment endpoint** (distinct from Pattern 2's batch *executions*), receiving instant insights to drive decisions—like checking risk scores before posting a sales order or providing recommendations as a page loads. + +**Triggers:** Real-time, synchronous calls triggered by user actions (e.g., "Submit," "Approve") or system processes (e.g., S/4HANA BAdI) requiring immediate responses. + +**Flow:** The application calls the AI Core deployment endpoint with event-specific records (e.g., line items from a sales order), receives predictions in milliseconds, and acts immediately—blocking transactions, flagging items for review, or displaying recommendations. + +**Why:** This pattern creates a **reusable AI asset**—the single model from Pattern 1 serves both massive batch jobs (Pattern 2) and critical real-time processes, embedding intelligence directly into enterprise operations without duplication. + +**Implementation with SAP Cloud Application Programming Model (CAP):** + +CAP provides a natural fit for implementing Pattern 3, offering significant advantages for deployments: + +* **Integrated Data Access:** CAP applications can seamlessly query SAP Datasphere (which federates BDC data products) and combine this with real-time AI predictions in a single request-response cycle, eliminating the need for separate data and inference layers. +* **Built-in Governance:** Authorization, authentication, and audit logging align automatically with SAP standards—the same security model protects both your data retrieval and AI inference calls. +* **Simplified Development:** Developers work within familiar SAP frameworks (CDS models, OData services) rather than managing low-level HTTP clients, reducing integration complexity and accelerating time-to-market. +* **Enterprise-Ready:** CAP applications deploy naturally into SAP BTP with built-in observability, scaling, and operational tooling—no additional infrastructure setup required. + +## Business Problem 1: AI-Enhanced Predictive Insights + +To make these patterns concrete, let's walk through a tangible, high-value example: **Improving Cash Flow with AI-Enhanced Payment Delay Predictions.** + +A large enterprise's finance department struggles with reactive cash flow management. They can see which payments are overdue, but they lack the foresight to act proactively. They need to not only *predict* which payments are likely to be delayed but also *understand why* so the collections team can prioritize their efforts and engage with customers in a more informed and targeted manner. + +### The Solution: Building an AI-Enhanced Data Product + +An AI-Enhanced Data Product is created by following the defined architectural patterns, with clear roles for each persona. + +**1. Model Development (Persona: Data Scientist in SAP BDC)** + +* The data scientist explores data in the SAP BDC catalog, identifying the `Entry View Journal Entry` data product. +* Working in an SAP Databricks notebook, they realize a simple prediction (the "what") is insufficient. They must also productionalize the explanation (the "why"). +* They prototype an **end-to-end prediction and explanation pipeline**. This pipeline: + 1. Trains an **XGBoost** model to get the prediction. + 2. Uses the **SHAP** library to calculate feature importance. + 3. Calls the **Generative AI Hub** (via the SAP AI SDK) to translate the SHAP values into a human-readable explanation (e.g., *"This payment is predicted to be 15 days late, primarily due to past payment behavior..."*). +* This entire, self-contained pipeline is saved as a single deployable asset. This completes the "Data Science on BDC" (development) part of **Pattern 1**. + +**2. Enterprise Deployment (Personas: ML Engineer & Data Scientist)** + +* This step follows the "MLOps on AI Core" (deployment) part of **Pattern 1: Train in Databricks, Serve in AI Core**. +* The **ML Engineer** *enables* this by first building a reusable serving template in **SAP AI Core** capable of running this complex (XGBoost + SHAP + GenAI) pipeline. +* The **Data Scientist**, still in their BDC notebook, uses the **SAP AI Core SDK** to trigger the deployment. They pass their saved pipeline asset to the ML Engineer's template. +* **SAP AI Core** automatically builds, containerizes, and deploys this entire pipeline, exposing a single, scalable, and governed API endpoint. This endpoint, when called, now returns the *full* payload: both the prediction and the explanation. + +**3. Operationalizing the "Why" (Personas: App Developer, Business User)** + +The deployed API is now operationalized using *both* consumption patterns to solve the business problem: + +* **Pattern 2: The Batch Consumption Pattern:** + * A **recurring workflow** is scheduled in **SAP AI Launchpad**. It follows this pattern to execute the native batch pipeline, reading all new invoices. + * It creates the **"Enriched Payment Forecasts" data product** in BDC, which now includes both the risk score *and* the human-readable explanation. + * The **Business User** opens their **SAP Analytics Cloud** dashboard, which reads this data product to see a fully-explained, prioritized worklist. + +* **Pattern 3: The Real-Time Consumption Pattern:** + * An **Application Developer** builds a Fiori app for the collections team. + * When a user opens a customer account, the app makes a *live call* to the *same* AI Core API for that customer's outstanding invoices. + * This pattern provides the team with instant, on-demand predictions and explanations to guide their conversation. + +### The Value Proposition (The "Why") + +This end-to-end scenario delivers value at multiple levels: + +- **For the Business:** The finance team moves from reactive to proactive, improving cash flow and enabling more meaningful customer interactions. The business gains a trusted, explainable AI solution, not a "black box." + +- **For the Data Scientist:** They can innovate rapidly in the agile Databricks environment while leveraging powerful, enterprise-grade AI capabilities from SAP AI Core without needing to be an expert in Kubernetes or API management. + +- **For IT & Governance:** The entire process is governed. Data access is controlled, the model is monitored, and the resulting data product is a managed asset. The architecture provides the robustness and auditability required for a mission-critical financial process. + +## Business Problem 2: Autonomous Process Optimization with AI Agents + +Beyond providing insights for users to act on, the architecture's true power is realized when the foundational patterns are combined to create **specialized AI data agents**. These are autonomous constructs that monitor, analyze, and act on business events with minimal human intervention, moving the business from a reactive to a proactive and automated posture. + +A practical example is a **"Cash Flow Optimization Agent"**, which builds directly on the previous scenario: + +1. **Monitor (Powered by Pattern 2):** The agent continuously observes the "Enriched Payment Forecasts" data product. This data is updated daily via **Pattern 2: The Batch Consumption Pattern**. + +2. **Analyze (Powered by Pattern 3):** When a high-risk invoice is detected, the agent *executes* **Pattern 3: The Real-Time Consumption Pattern**. It makes a real-time call to the deployed API (the one created by Pattern 1) to get an immediate, deep explanation. + +3. **Act (The Next Step):** Based on the analysis from Pattern 3, the agent triggers a workflow in SAP BTP to assign a task or uses a generative AI model to draft a personalized outreach email. + +### Key Recommendations and Best Practices for Batch Inference + +* **Utilize Workflow Executions:** Orchestrate batch inference as pipelines/workflow executions in SAP AI Core to manage high-volume, offline processing efficiently. +* **Scheduling:** Schedule recurring batch jobs using cron specifications in SAP AI Launchpad to automate periodic tasks (e.g., weekly or monthly reports). +* **Efficiency:** Prioritize throughput over latency; leverage GPUs and other compute resources efficiently for large datasets. +* **Data Handling:** Design batch programs to process data in bulk, reading from and writing to configured object stores (e.g., S3). Produce an output dataset of predictions for downstream consumption or as a governed data product. +* **Containerization:** Package batch inference logic in a user-supplied Docker image containing the necessary code and dependencies (e.g., PyTorch, TensorFlow) for execution in SAP AI Core. +* **API-First Approach:** Manage and trigger configurations and executions via SAP AI Core APIs or the Python SDK to integrate with external applications and pipelines. +* **Separate from Online Serving:** Do not use real-time serving endpoints for batch; online endpoints are optimized for low-latency single or small-batch requests, whereas workflow executions are designed for high-volume, offline processing. +* **Resource Management:** Leverage platform capabilities to dynamically scale compute resources based on job demands to ensure cost-efficiency and performance. + +## Summary of Roles and Responsibilities + +- **Data Scientists:** Primarily work within **SAP Business Data Cloud (Databricks)** for data exploration, feature engineering, and rapid model experimentation. +- **ML Engineers & IT Operations:** Primarily work with **SAP AI Core** to manage production model deployments, monitor performance, ensure governance, and maintain the operational integrity of AI services. +- **Application Developers & Business Users:** Consume the final AI-Enhanced Data Products and AI-powered applications through various channels, including SAP Analytics Cloud, custom BTP applications, or integrated line-of-business solutions. + +## Components and Further Reading + +This reference architecture is realized through the following key SAP services: +- SAP Business Data Cloud +- SAP AI Core +- SAP Generative AI Hub +- SAP Databricks +- SAP Datasphere +- SAP Analytics Cloud + +For detailed implementation guides and "how-to" tutorials, please refer to the official SAP documentation and related technical blog posts. From 301ed9bdbfbec6172f5f347617ce9eca07d7c7f7 Mon Sep 17 00:00:00 2001 From: I321968 Date: Wed, 10 Dec 2025 15:19:05 -0800 Subject: [PATCH 05/19] Fix broken site, add diagram --- docs/learning jorney/Unit4.md | 18 +- .../drawio/BDC_AI_External_Arch 1.drawio | 4480 +++++++++++++++++ .../drawio/bdc-ai-core-integration.drawio | 4477 ++++++++++++++++ .../RA0013/7-bdc-powered-by-ai-core/readme.md | 30 +- 4 files changed, 8993 insertions(+), 12 deletions(-) create mode 100644 docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio create mode 100644 docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio diff --git a/docs/learning jorney/Unit4.md b/docs/learning jorney/Unit4.md index 86d64a52fb..d174515f95 100644 --- a/docs/learning jorney/Unit4.md +++ b/docs/learning jorney/Unit4.md @@ -207,7 +207,7 @@ When evaluating a BDC architecture, consider these five critical areas: | Decision | Option A | Option B | Choose A When... | Choose B When... | |----------|----------|----------|------------------|------------------| | Data Access | Federation | Replication | Data freshness critical, storage expensive | Query performance critical, budget allows | -| Processing | Real-time | Batch | Latency <1 sec required | Cost optimization priority, can tolerate delay | +| Processing | Real-time | Batch | Latency <1 sec required | Cost optimization priority, can tolerate delay | | Data Products | SAP-managed | Customer-managed | Standard use cases, faster deployment | Custom logic, full control needed | **Reference:** [RA0013: SAP Business Data Cloud](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6) @@ -222,15 +222,15 @@ When evaluating a BDC architecture, consider these five critical areas: **Strategic Deployment Patterns:** -1. **Greenfield (New Implementation)** - - Use IAS as primary identity provider - - Simpler setup, faster deployment - - Best for: SMEs, new SAP landscape +**Greenfield (New Implementation):** +- Use IAS as primary identity provider +- Simpler setup, faster deployment +- Best for: SMEs, new SAP landscape -2. **Brownfield (Existing Enterprise IdP)** - - Federate enterprise IdP (Azure AD, Okta) with IAS - - IAS acts as broker to SAP systems - - Best for: Large enterprises, existing IAM infrastructure +**Brownfield (Existing Enterprise IdP):** +- Federate enterprise IdP (Azure AD, Okta) with IAS +- IAS acts as broker to SAP systems +- Best for: Large enterprises, existing IAM infrastructure **Zero-Trust Principles:** - Multi-factor authentication (MFA) mandatory diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio new file mode 100644 index 0000000000..db2fdc7855 --- /dev/null +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio @@ -0,0 +1,4480 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio new file mode 100644 index 0000000000..8b7f4f4083 --- /dev/null +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio @@ -0,0 +1,4477 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index a9840a4dda..984a0a922b 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -32,19 +32,21 @@ toc_max_heading_level: 4 draft: false unlisted: false contributors: - - SeeObjective + - seeobjectively - guilherme-segantini - jmsrpp - anbazhagan-uma last_update: - author: guilherme-segantini - date: 2025-12-05 + author: seeobjectively + date: 2025-12-10 --- Enterprises possess a wealth of invaluable business data within their SAP systems. However, activating this data for modern Artificial Intelligence is often a complex, disconnected, and risky endeavor. To stay competitive, organizations need a strategy to transform this data into reliable, governed, and actionable AI-driven insights that are deeply integrated with core business processes. This reference architecture presents a cohesive vision for combining **SAP Business Data Cloud** with **SAP AI Foundation** (including SAP AI Core and the Generative AI Hub). The core architectural concept is the creation of **AI-Enhanced Data Products**—intelligent, context-aware, and dynamic assets that deliver trusted predictive insights and drive business automation at scale. This integrated approach enables the development of AI solutions that are **reliable, relevant, and responsible**, accelerating time-to-value and embedding intelligence directly into the enterprise. +![drawio](drawio/bdc-ai-core-integration.drawio) + ## The Architectural Blueprint The architectural vision is centered on a powerful synergy: SAP Business Data Cloud provides the governed, business-ready data foundation, while SAP AI Foundation offers a comprehensive portfolio of enterprise-grade AI capabilities. This avoids a "one-size-fits-all" approach, recognizing that different business problems require different tools. The value lies in the comphreensive AI capabilities that allows data scientists and developers to move from data preparation to model deployment and inference. @@ -189,6 +191,28 @@ A practical example is a **"Cash Flow Optimization Agent"**, which builds direct - **ML Engineers & IT Operations:** Primarily work with **SAP AI Core** to manage production model deployments, monitor performance, ensure governance, and maintain the operational integrity of AI services. - **Application Developers & Business Users:** Consume the final AI-Enhanced Data Products and AI-powered applications through various channels, including SAP Analytics Cloud, custom BTP applications, or integrated line-of-business solutions. +## Platform Selection Guide + +### Use **Databricks** for: +- **Data science experimentation** and rapid prototyping +- **Quick validation** by data scientist personas +- Integrated data science workflows with immediate feedback + +### Use **AI Core** for: +- **Production-ready models** requiring enterprise-grade serving +- Access to **broader LLM ecosystem** and **SAP's RPT-1** foundational model +- Scenarios requiring **production infrastructure** and planned **BDC integrations** + +## Key Differentiators + +| Aspect | AI Core | Databricks | +|--------|---------|------------| +| **Use Case** | Production deployment | Experimentation & prototyping | +| **Infrastructure** | Enterprise-grade serving | Integrated development environment | +| **Model Access** | Broad LLM ecosystem + RPT-1 | Selected LLMs | +| **Speed to Value** | Production-ready deployment | Rapid prototyping | +| **Integration** | Planned BDC integrations | Native data science workflows | + ## Components and Further Reading This reference architecture is realized through the following key SAP services: From 126edc78f2d56187c83904b3ef60f92eca55f9ed Mon Sep 17 00:00:00 2001 From: I321968 Date: Wed, 10 Dec 2025 15:26:52 -0800 Subject: [PATCH 06/19] Remove unused code --- docs/learning jorney/Unit4.md | 363 ---------------------------------- 1 file changed, 363 deletions(-) delete mode 100644 docs/learning jorney/Unit4.md diff --git a/docs/learning jorney/Unit4.md b/docs/learning jorney/Unit4.md deleted file mode 100644 index d174515f95..0000000000 --- a/docs/learning jorney/Unit4.md +++ /dev/null @@ -1,363 +0,0 @@ -# Designing Hybrid and Customized Architectures for Enterprise Architects -## Strategic Foundations Course - -## Course Overview - -This entry-level strategic course equips enterprise architects with fundamental knowledge to design hybrid architectures that integrate SAP and non-SAP systems, leverage AI capabilities, and ensure data reliability. Participants will learn to apply SAP Architecture Center reference architectures as practical blueprints for customer implementations, focusing on strategic decision-making and architectural assessment. - -**Target Audience:** Enterprise Architects (entry level) -**Duration:** 90-120 minutes -**Focus:** Strategic fundamentals, not deep technical implementation - ---- - -## Course Structure - -### Lesson 1: Hybrid Architecture Strategy (30 minutes) - -#### Learning Objectives: -By the end of this lesson, participants will be able to: -- Explain the strategic value of hybrid architectures for enterprise transformation -- Identify when to use cloud, on-premise, and multi-cloud patterns -- Understand SAP Business Data Cloud as a unified data platform -- Leverage SAP Architecture Center reference architectures for decision-making - -#### 1.1 Why Hybrid Architectures Matter (10 minutes) - -**Strategic Drivers:** -- **Business Agility:** Rapid deployment of new capabilities without replacing legacy systems -- **Data Sovereignty:** Compliance with regional regulations (GDPR, data residency) -- **Cost Optimization:** Balance cloud economics with existing infrastructure investments -- **Risk Mitigation:** Gradual migration reduces disruption - -**Key Decision Points:** -- Which workloads belong in cloud vs. on-premise? -- How to maintain data consistency across environments? -- What integration patterns minimize complexity? - -**SAP Architecture Center Value:** -- Pre-validated blueprints reduce architecture risk -- Standardized patterns accelerate time-to-value -- Alignment with SAP's strategic roadmap - -**Reference:** [RA0008: Edge Integration Cell](https://architecture.learning.sap.com/docs/ref-arch/263f576c90) - -#### 1.2 SAP Business Data Cloud - Strategic Overview (15 minutes) - -**What is SAP BDC?** -A unified platform that brings together SAP and non-SAP data for analytics and AI: -- **SAP Datasphere:** Connect and harmonize data from any source -- **SAP Analytics Cloud:** Business intelligence and planning -- **SAP Databricks:** Advanced AI/ML capabilities - -**Strategic Capabilities:** -- **Data Products:** Reusable, governed data assets -- **Federation vs. Replication:** Choose based on latency needs and cost -- **Zero-Copy Architecture:** Access data without moving it (via Delta Share) - -**When to Use BDC:** -- ✓ Need unified view across SAP and non-SAP systems -- ✓ Building AI/ML solutions requiring high-quality data -- ✓ Modernizing analytics infrastructure -- ✗ Simple reporting from single SAP system (consider native tools first) - -**Reference:** [RA0013: SAP Business Data Cloud](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6) - -#### 1.3 Integration and Security Fundamentals (5 minutes) - -**Integration Strategy:** -- **API-First:** SAP Integration Suite for orchestration -- **Event-Driven:** Real-time data flows for critical processes -- **Hybrid Execution:** Edge Integration Cell for on-premise connectivity - -**Security Essentials:** -- Zero-trust principles (verify every access) -- Identity and access management (covered in Lesson 3) -- Secure connectivity (Cloud Connector, Private Link) - -**Reference:** [RA0001: Event-Driven Architecture](https://architecture.learning.sap.com/docs/ref-arch/fbdc46aaae) - ---- - -### Lesson 2: AI Agent Strategy and Interoperability (30 minutes) - -#### Learning Objectives: -By the end of this lesson, participants will be able to: -- Understand when and why to use AI agents in enterprise architectures -- Distinguish between SAP's two interoperability standards: MCP and A2A -- Apply Agent2Agent (A2A) protocol for cross-platform agent collaboration -- Make strategic decisions about agent deployment in hybrid landscapes - -#### 2.1 AI Agents - Strategic Perspective (10 minutes) - -**What are AI Agents?** -Autonomous systems that can reason, plan, and execute multi-step tasks with minimal human intervention. - -**When to Consider AI Agents:** -- ✓ Complex workflows spanning multiple systems -- ✓ Tasks requiring contextual decision-making -- ✓ Processes benefiting from natural language interaction -- ✗ Simple automation (use traditional workflow tools) -- ✗ Single-step operations (use standard APIs) - -**Two Development Paths:** -1. **Content-Based (Joule Studio):** Low-code, business-user friendly, rapid deployment -2. **Code-Based (LangGraph, AutoGen):** Full control, complex custom logic, developer-led - -**Strategic Question:** Which approach fits your organization's skills and timeline? - -**Reference:** [RA0005: AI Agents & Agent Builder](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d/5) - -#### 2.2 Interoperability Strategy: MCP vs. A2A (15 minutes) - -**Why Interoperability Matters:** -In hybrid landscapes, SAP Joule agents need to collaborate with third-party agents (Microsoft Copilot, Google Vertex AI, AWS Bedrock). SAP's strategy uses two standards with distinct purposes: - -**1. Model Context Protocol (MCP) - Internal Use Only** -- **Purpose:** Connect AI agents to SAP data and tools -- **Scope:** Internal to SAP systems -- **Key Point:** Not for customer architecture design - SAP manages MCP internally - -**2. Agent-to-Agent (A2A) - External Collaboration** -- **Purpose:** Enable agents from different vendors to work together -- **Scope:** Cross-platform, vendor-neutral standard -- **Key Point:** This is what enterprise architects should design with - -**Strategic Architecture Principle:** -> Use A2A as the standard pattern for agent collaboration in hybrid architectures. Ensure SAP remains the governed data hub, with external agents interacting via A2A protocol only. - -**Why A2A Matters:** -- **Security:** Agents collaborate without exposing internal data access -- **Governance:** Centralized control through Agent & Tool Gateway -- **Flexibility:** Plug-and-play ecosystem of specialized agents -- **Standardization:** Vendor-neutral, future-proof approach - -**Architect's Decision Framework:** -- Third-party agent integration? → Use A2A protocol -- Data access from agents? → Route through SAP BDC with governance -- Custom agent development? → Consider Joule Studio first, code-based if needed - -**Reference:** [RA0005: Agent Interoperability](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d/8) - -#### 2.3 Data Access Strategy for Agents (5 minutes) - -**Two Primary Data Access Patterns:** - -1. **Structured Data (SQL-based)** - - Agents query SAP Datasphere using natural language - - Federated access across SAP and non-SAP sources - - Role-based security enforced - -2. **Unstructured Data (RAG - Retrieval Augmented Generation)** - - Vector search in SAP HANA Cloud for document similarity - - Grounding AI responses in enterprise knowledge - - Reduces hallucination, improves accuracy - -**Governance Essentials:** -- Data lineage tracking -- Access audit trails -- Compliance (GDPR, industry regulations) - -**Reference:** [RA0005: Retrieval Augmented Generation](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d/3) - ---- - -### Lesson 3: Assessing SAP Business Data Cloud Architecture and Engineering for Increased AI Reliability (30 minutes) - -#### Learning Objectives: -By the end of this lesson, participants will be able to: -- Assess SAP Business Data Cloud architectures using a structured framework -- Understand identity and access management strategy for BDC -- Apply key reliability principles for AI systems -- Make informed architecture decisions based on trade-off analysis - -#### 3.1 BDC Architecture Assessment Framework (15 minutes) - -**Strategic Assessment Dimensions:** - -When evaluating a BDC architecture, consider these five critical areas: - -1. **Scalability** - - Can it handle 10x data growth? - - Concurrent user capacity adequate? - - Strategy: Federation (flexible) vs. Replication (performance) - -2. **Security** - - Zero-trust implementation (verify every access) - - Data encryption (at rest and in transit) - - Identity and access management maturity - -3. **Reliability** - - Uptime targets (99.9% = 8.7 hours downtime/year) - - Data freshness requirements (real-time vs. batch) - - Disaster recovery plan (RPO/RTO defined?) - -4. **Cost Optimization** - - Federation saves storage, but may increase compute - - Replication costs more storage, but faster queries - - Data lifecycle policies to archive cold data - -5. **AI/ML Readiness** - - Data quality sufficient for model training? - - Feature engineering capabilities available? - - Production deployment infrastructure ready? - -**Key Trade-off Decisions:** - -| Decision | Option A | Option B | Choose A When... | Choose B When... | -|----------|----------|----------|------------------|------------------| -| Data Access | Federation | Replication | Data freshness critical, storage expensive | Query performance critical, budget allows | -| Processing | Real-time | Batch | Latency <1 sec required | Cost optimization priority, can tolerate delay | -| Data Products | SAP-managed | Customer-managed | Standard use cases, faster deployment | Custom logic, full control needed | - -**Reference:** [RA0013: SAP Business Data Cloud](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6) - -#### 3.2 Identity and Access Management Strategy (10 minutes) - -**SAP Cloud Identity Services (CIS):** - -**Core Components:** -- **IAS (Identity Authentication):** Single sign-on across SAP and BDC -- **IPS (Identity Provisioning):** Automated user lifecycle management - -**Strategic Deployment Patterns:** - -**Greenfield (New Implementation):** -- Use IAS as primary identity provider -- Simpler setup, faster deployment -- Best for: SMEs, new SAP landscape - -**Brownfield (Existing Enterprise IdP):** -- Federate enterprise IdP (Azure AD, Okta) with IAS -- IAS acts as broker to SAP systems -- Best for: Large enterprises, existing IAM infrastructure - -**Zero-Trust Principles:** -- Multi-factor authentication (MFA) mandatory -- Least privilege access (only what's needed) -- Continuous validation (not "trust once") -- Centralized audit logging - -**Why This Matters for AI:** -- AI agents need controlled access to data -- Granular permissions prevent data leakage -- Audit trails for compliance and debugging - -**Reference:** [RA0019: Identity and Access Management](https://architecture.learning.sap.com/docs/ref-arch/20c6b29b1e) - -#### 3.3 AI Reliability Essentials (5 minutes) - -**Key Reliability Principles:** - -1. **Data Quality** - - Automated validation at ingestion - - Monitor for drift (data distribution changes) - - Schema evolution management - -2. **Model Lifecycle** - - Separate environments (dev, staging, prod) - - Gradual rollout (canary deployments) - - Automated retraining when accuracy drops - -3. **Monitoring** - - Track: Latency, accuracy, throughput - - Alert on: Performance degradation, errors - - Review: Regular postmortem analysis - -4. **Governance** - - Model approval workflows - - Bias detection and fairness metrics - - Compliance (GDPR, regulatory requirements) - -**Critical Questions for Architects:** -- What's the acceptable downtime for AI services? -- How quickly must we detect model degradation? -- Who approves model deployments to production? -- How do we ensure AI decisions are explainable? - -**Reference:** [RA0005: Generative AI on SAP BTP](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d) - ---- - -## Key SAP Architecture Center Reference Architectures - -### Essential References: -- **[RA0001: Event-Driven Architecture (EDA)](https://architecture.learning.sap.com/docs/ref-arch/fbdc46aaae)** -- **[RA0005: Generative AI on SAP BTP](https://architecture.learning.sap.com/docs/ref-arch/e5eb3b9b1d)** -- **[RA0008: Edge Integration Cell](https://architecture.learning.sap.com/docs/ref-arch/263f576c90)** -- **[RA0013: SAP Business Data Cloud](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6)** (Primary Reference) -- **[RA0019: Identity and Access Management](https://architecture.learning.sap.com/docs/ref-arch/20c6b29b1e)** -- **[RA0024: Integrating and Extending Joule](https://architecture.learning.sap.com/docs/ref-arch/06ff6062dc)** - -### Supporting References: -- **[RA0004: Modernizing SAP BW with SAP BDC](https://architecture.learning.sap.com/docs/ref-arch/f5b6b597a6/4)** -- **[RA0012: HANA Cloud Medallion Architecture](https://architecture.learning.sap.com/docs/ref-arch/d9b25daf96)** -- **[RA0026: Embodied AI Agents](https://architecture.learning.sap.com/docs/ref-arch/083f2d968e)** - ---- - -## Course Takeaways - -### What You've Learned: - -1. **Hybrid Architecture Design** - - Integration patterns for cloud, on-premise, and multi-cloud - - Data federation vs. replication strategies - - Secure connectivity and zero-trust principles - -2. **AI Agent Development & Interoperability** - - Content-based vs. code-based development approaches - - Agent2Agent protocol for cross-platform collaboration - - Data access strategies for structured and unstructured data - -3. **SAP BDC for AI Reliability** - - Unified data platform architecture - - Identity and access management with Cloud Identity Services - - AI governance and compliance frameworks - -### How to Apply SAP Architecture Center: - -- **Start with Reference Architectures:** Use proven blueprints as starting points -- **Adapt to Context:** Customize patterns for your specific requirements -- **Follow Best Practices:** Leverage SAP-recommended approaches -- **Stay Current:** Monitor updates to reference architectures -- **Engage Community:** Contribute and learn from peer implementations - ---- - -## Next Steps - -### Immediate Actions: -1. Explore SAP Architecture Center for your specific use cases -2. Identify hybrid integration requirements in your landscape -3. Assess AI agent opportunities for automation -4. Review identity and access management maturity - -### Advanced Learning: -- Deep-dive workshops on specific reference architectures -- Hands-on labs with SAP BTP services -- Architecture review sessions with SAP experts -- Implementation guidance and best practices - -### Resources: -- SAP Architecture Center: Comprehensive reference architectures -- SAP Discovery Center: Guided missions and tutorials -- SAP Help Documentation: Technical specifications -- SAP Community: Blogs, forums, and peer insights - ---- - -## Additional Resources - -### Documentation: -- [SAP Architecture Center](https://www.sap.com/architecture-center) -- [SAP Business Data Cloud](https://www.sap.com/products/technology-platform/business-data-cloud.html) -- [SAP Cloud Identity Services](https://discovery-center.cloud.sap/serviceCatalog/cloud-identity-services) -- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) - -### Sample Implementations: -- GitHub: SAP-samples repositories -- SAP Discovery Center: Mission-based learning paths -- SAP Community: Blog posts and tutorials - ---- From 197251965e69d825427bb2f13c11c5b40751d8ba Mon Sep 17 00:00:00 2001 From: I321968 Date: Wed, 10 Dec 2025 15:30:17 -0800 Subject: [PATCH 07/19] Remove duplicate diagram --- .../drawio/BDC_AI_External_Arch 1.drawio | 4480 ----------------- 1 file changed, 4480 deletions(-) delete mode 100644 docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio deleted file mode 100644 index db2fdc7855..0000000000 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/BDC_AI_External_Arch 1.drawio +++ /dev/null @@ -1,4480 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From d2797b11a2b9c7f6468fa22f964d1a2078d7f4b1 Mon Sep 17 00:00:00 2001 From: I321968 Date: Thu, 11 Dec 2025 09:17:42 -0800 Subject: [PATCH 08/19] Update diagram and clean up readme --- .../drawio/bdc-ai-core-integration.drawio | 112 ++++++++++-------- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 14 +-- 2 files changed, 62 insertions(+), 64 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio index 8b7f4f4083..ba1b001843 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio @@ -1,6 +1,6 @@ - + @@ -46,22 +46,27 @@ - + - + - + - - + + + + + + + - + @@ -69,10 +74,10 @@ - + - + @@ -113,7 +118,7 @@ - + @@ -122,7 +127,7 @@ - + @@ -210,19 +215,16 @@ - + - + - - - - + @@ -233,7 +235,7 @@ - + @@ -250,7 +252,7 @@ - + @@ -259,54 +261,62 @@ - - - - - - - - - - - - - - - - - - - - + - + - + - - + + + + + - - + + - - + + - + - - - + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 984a0a922b..f12fc66f38 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -191,25 +191,13 @@ A practical example is a **"Cash Flow Optimization Agent"**, which builds direct - **ML Engineers & IT Operations:** Primarily work with **SAP AI Core** to manage production model deployments, monitor performance, ensure governance, and maintain the operational integrity of AI services. - **Application Developers & Business Users:** Consume the final AI-Enhanced Data Products and AI-powered applications through various channels, including SAP Analytics Cloud, custom BTP applications, or integrated line-of-business solutions. -## Platform Selection Guide - -### Use **Databricks** for: -- **Data science experimentation** and rapid prototyping -- **Quick validation** by data scientist personas -- Integrated data science workflows with immediate feedback - -### Use **AI Core** for: -- **Production-ready models** requiring enterprise-grade serving -- Access to **broader LLM ecosystem** and **SAP's RPT-1** foundational model -- Scenarios requiring **production infrastructure** and planned **BDC integrations** - ## Key Differentiators | Aspect | AI Core | Databricks | |--------|---------|------------| | **Use Case** | Production deployment | Experimentation & prototyping | | **Infrastructure** | Enterprise-grade serving | Integrated development environment | -| **Model Access** | Broad LLM ecosystem + RPT-1 | Selected LLMs | +| **Model Access** | Broad LLM ecosystem & SAP's foundation models like RPT-1 | Selected LLMs | | **Speed to Value** | Production-ready deployment | Rapid prototyping | | **Integration** | Planned BDC integrations | Native data science workflows | From f88149eea73e0c3f7f3b8659af36b3f7cbdb71e1 Mon Sep 17 00:00:00 2001 From: I321968 Date: Thu, 11 Dec 2025 09:31:48 -0800 Subject: [PATCH 09/19] update tags, add links to related content --- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 53 ++++++++++++++----- 1 file changed, 40 insertions(+), 13 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index f12fc66f38..916dcfbf26 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -17,14 +17,11 @@ keywords: - reference architecture sidebar_label: SAP Business Data Cloud powered by SAP AI Core tags: - - ai - data - - sap-ai-core - genai - databricks - - mlops - - architecture - - data-products + - bdc + - agents hide_table_of_contents: false hide_title: false toc_min_heading_level: 2 @@ -203,12 +200,42 @@ A practical example is a **"Cash Flow Optimization Agent"**, which builds direct ## Components and Further Reading -This reference architecture is realized through the following key SAP services: -- SAP Business Data Cloud -- SAP AI Core -- SAP Generative AI Hub -- SAP Databricks -- SAP Datasphere -- SAP Analytics Cloud +This reference architecture is realized through the following key SAP services and components: -For detailed implementation guides and "how-to" tutorials, please refer to the official SAP documentation and related technical blog posts. +### Related Reference Architectures + +**SAP Business Data Cloud Series:** +- [Data Products in SAP Business Data Cloud](../1-data-products-in-sap-business-data-cloud/readme.md) - Understanding data products, their architecture, and consumption patterns +- [SAP Databricks in SAP BDC](../5-sap-databricks-in-business-data-cloud/readme.md) - Deep dive into SAP Databricks integration and use cases +- [Intelligent Applications in SAP Business Data Cloud](../2-intelligent-applications-by-sap/readme.md) - Pre-configured analytics and dashboards +- [Modernizing SAP BW with SAP Business Data Cloud](../4-modernizing-sap-bw-with-sap-bdc/readme.md) - Migration patterns and data product generation +- [Cloud Identity Services for BDC](../6-cloud-identity-services-bdc/readme.md) - Unified identity and access management + +**Generative AI and Machine Learning:** +- [Generative AI with SAP AI Core](../../RA0005/readme.md) - Comprehensive guide to GenAI patterns, RAG, and AI agents +- [Federated Machine Learning with SAP Datasphere](../../RA0003/readme.md) - ML integration across hyperscaler platforms + +### SAP Services and Documentation + +**SAP AI Foundation:** +- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) - Enterprise AI platform for model lifecycle management +- [SAP AI Launchpad](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-launchpad) - Multi-tenant SaaS for managing AI scenarios +- [Generative AI Hub in SAP AI Core](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/generative-ai-hub-in-sap-ai-core) - Access to foundation models and LLMs +- [SAP AI Core SDK](https://pypi.org/project/ai-core-sdk/) - Python SDK for programmatic AI Core integration + +**SAP Business Data Cloud:** +- [SAP Business Data Cloud Overview](https://www.sap.com/products/technology-platform/business-data-cloud.html) - Product overview and capabilities +- [SAP Datasphere](https://help.sap.com/docs/SAP_DATASPHERE) - Data management, modeling, and integration +- [SAP Analytics Cloud](https://www.sap.com/products/technology-platform/cloud-analytics.html) - Business intelligence and analytics +- [SAP Databricks Documentation](https://help.sap.com/docs/sap-datasphere/sap-datasphere-administration-guide-for-sap-datasphere/sap-databricks) - Integration guide for SAP Databricks + +**Development and Integration:** +- [SAP Cloud Application Programming Model (CAP)](https://cap.cloud.sap/docs/) - Framework for building enterprise applications +- [SAP AI SDK for JavaScript/TypeScript](https://github.com/SAP/ai-sdk-js) - SDK for integrating AI capabilities +- [SAP BDC Connect SDK](https://pypi.org/project/sap-bdc-connect-sdk/) - Python SDK for data product management + +### Learning Resources + +- [Introducing SAP Business Data Cloud](https://learning.sap.com/learning-journeys/introducing-sap-business-data-cloud) - Learning journey +- [SAP AI Core Tutorial](https://developers.sap.com/tutorials/ai-core-genaihub-provisioning.html) - Set up Generative AI Hub +- [SAP Community: Business Data Cloud](https://community.sap.com/topics/business-data-cloud) - Community discussions and blogs From cb665b25480e5691bda95c2fe55d6d617f3da5ee Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Thu, 11 Dec 2025 10:25:47 -0800 Subject: [PATCH 10/19] Add mcp content --- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 73 ++++++++----------- 1 file changed, 31 insertions(+), 42 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 916dcfbf26..b3886c8742 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -17,11 +17,14 @@ keywords: - reference architecture sidebar_label: SAP Business Data Cloud powered by SAP AI Core tags: + - ai - data + - sap-ai-core - genai - databricks - - bdc - - agents + - mlops + - architecture + - data-products hide_table_of_contents: false hide_title: false toc_min_heading_level: 2 @@ -42,6 +45,10 @@ Enterprises possess a wealth of invaluable business data within their SAP system This reference architecture presents a cohesive vision for combining **SAP Business Data Cloud** with **SAP AI Foundation** (including SAP AI Core and the Generative AI Hub). The core architectural concept is the creation of **AI-Enhanced Data Products**—intelligent, context-aware, and dynamic assets that deliver trusted predictive insights and drive business automation at scale. This integrated approach enables the development of AI solutions that are **reliable, relevant, and responsible**, accelerating time-to-value and embedding intelligence directly into the enterprise. +**Future State: Real-Time Data Access via MCP** + +To further enhance real-time consumption patterns, SAP is developing a managed Model Context Protocol (MCP) component that will provide a standardized way to query derived data products directly from SAP Business Data Cloud. This future capability will simplify real-time data access for AI applications, enabling seamless integration of governed BDC data products into live business processes. + ![drawio](drawio/bdc-ai-core-integration.drawio) ## The Architectural Blueprint @@ -103,7 +110,7 @@ The strategy we're outlining provides a clear, governed path to solve this. It's CAP provides a natural fit for implementing Pattern 3, offering significant advantages for deployments: -* **Integrated Data Access:** CAP applications can seamlessly query SAP Datasphere (which federates BDC data products) and combine this with real-time AI predictions in a single request-response cycle, eliminating the need for separate data and inference layers. +* **Integrated Data Access:** CAP applications can seamlessly query SAP Datasphere (which federates BDC data products) and combine this with real-time AI predictions in a single request-response cycle, eliminating the need for separate data and inference layers. In the future, the SAP-managed MCP component will further streamline data retrieval by providing a standardized query interface for BDC data products. Applications can then combine this contextual business data with AI Core predictions (via separate API calls) to create a complete real-time intelligence flow. * **Built-in Governance:** Authorization, authentication, and audit logging align automatically with SAP standards—the same security model protects both your data retrieval and AI inference calls. * **Simplified Development:** Developers work within familiar SAP frameworks (CDS models, OData services) rather than managing low-level HTTP clients, reducing integration complexity and accelerating time-to-market. * **Enterprise-Ready:** CAP applications deploy naturally into SAP BTP with built-in observability, scaling, and operational tooling—no additional infrastructure setup required. @@ -188,54 +195,36 @@ A practical example is a **"Cash Flow Optimization Agent"**, which builds direct - **ML Engineers & IT Operations:** Primarily work with **SAP AI Core** to manage production model deployments, monitor performance, ensure governance, and maintain the operational integrity of AI services. - **Application Developers & Business Users:** Consume the final AI-Enhanced Data Products and AI-powered applications through various channels, including SAP Analytics Cloud, custom BTP applications, or integrated line-of-business solutions. +## Platform Selection Guide + +### Use **Databricks** for: +- **Data science experimentation** and rapid prototyping +- **Quick validation** by data scientist personas +- Integrated data science workflows with immediate feedback + +### Use **AI Core** for: +- **Production-ready models** requiring enterprise-grade serving +- Access to **broader LLM ecosystem** and **SAP's RPT-1** foundational model +- Scenarios requiring **production infrastructure** and planned **BDC integrations** + ## Key Differentiators | Aspect | AI Core | Databricks | |--------|---------|------------| | **Use Case** | Production deployment | Experimentation & prototyping | | **Infrastructure** | Enterprise-grade serving | Integrated development environment | -| **Model Access** | Broad LLM ecosystem & SAP's foundation models like RPT-1 | Selected LLMs | +| **Model Access** | Broad LLM ecosystem + RPT-1 | Selected LLMs | | **Speed to Value** | Production-ready deployment | Rapid prototyping | | **Integration** | Planned BDC integrations | Native data science workflows | ## Components and Further Reading -This reference architecture is realized through the following key SAP services and components: - -### Related Reference Architectures - -**SAP Business Data Cloud Series:** -- [Data Products in SAP Business Data Cloud](../1-data-products-in-sap-business-data-cloud/readme.md) - Understanding data products, their architecture, and consumption patterns -- [SAP Databricks in SAP BDC](../5-sap-databricks-in-business-data-cloud/readme.md) - Deep dive into SAP Databricks integration and use cases -- [Intelligent Applications in SAP Business Data Cloud](../2-intelligent-applications-by-sap/readme.md) - Pre-configured analytics and dashboards -- [Modernizing SAP BW with SAP Business Data Cloud](../4-modernizing-sap-bw-with-sap-bdc/readme.md) - Migration patterns and data product generation -- [Cloud Identity Services for BDC](../6-cloud-identity-services-bdc/readme.md) - Unified identity and access management - -**Generative AI and Machine Learning:** -- [Generative AI with SAP AI Core](../../RA0005/readme.md) - Comprehensive guide to GenAI patterns, RAG, and AI agents -- [Federated Machine Learning with SAP Datasphere](../../RA0003/readme.md) - ML integration across hyperscaler platforms - -### SAP Services and Documentation - -**SAP AI Foundation:** -- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) - Enterprise AI platform for model lifecycle management -- [SAP AI Launchpad](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-launchpad) - Multi-tenant SaaS for managing AI scenarios -- [Generative AI Hub in SAP AI Core](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/generative-ai-hub-in-sap-ai-core) - Access to foundation models and LLMs -- [SAP AI Core SDK](https://pypi.org/project/ai-core-sdk/) - Python SDK for programmatic AI Core integration - -**SAP Business Data Cloud:** -- [SAP Business Data Cloud Overview](https://www.sap.com/products/technology-platform/business-data-cloud.html) - Product overview and capabilities -- [SAP Datasphere](https://help.sap.com/docs/SAP_DATASPHERE) - Data management, modeling, and integration -- [SAP Analytics Cloud](https://www.sap.com/products/technology-platform/cloud-analytics.html) - Business intelligence and analytics -- [SAP Databricks Documentation](https://help.sap.com/docs/sap-datasphere/sap-datasphere-administration-guide-for-sap-datasphere/sap-databricks) - Integration guide for SAP Databricks - -**Development and Integration:** -- [SAP Cloud Application Programming Model (CAP)](https://cap.cloud.sap/docs/) - Framework for building enterprise applications -- [SAP AI SDK for JavaScript/TypeScript](https://github.com/SAP/ai-sdk-js) - SDK for integrating AI capabilities -- [SAP BDC Connect SDK](https://pypi.org/project/sap-bdc-connect-sdk/) - Python SDK for data product management - -### Learning Resources +This reference architecture is realized through the following key SAP services: +- SAP Business Data Cloud +- SAP AI Core +- SAP Generative AI Hub +- SAP Databricks +- SAP Datasphere +- SAP Analytics Cloud -- [Introducing SAP Business Data Cloud](https://learning.sap.com/learning-journeys/introducing-sap-business-data-cloud) - Learning journey -- [SAP AI Core Tutorial](https://developers.sap.com/tutorials/ai-core-genaihub-provisioning.html) - Set up Generative AI Hub -- [SAP Community: Business Data Cloud](https://community.sap.com/topics/business-data-cloud) - Community discussions and blogs +For detailed implementation guides and "how-to" tutorials, please refer to the official SAP documentation and related technical blog posts. \ No newline at end of file From c3855cad20e927fb1c8212c71850e793543c6ab0 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Thu, 11 Dec 2025 10:42:17 -0800 Subject: [PATCH 11/19] Fix references section --- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 49 +++++++++++++++---- 1 file changed, 40 insertions(+), 9 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index b3886c8742..1243f69b6c 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -219,12 +219,43 @@ A practical example is a **"Cash Flow Optimization Agent"**, which builds direct ## Components and Further Reading -This reference architecture is realized through the following key SAP services: -- SAP Business Data Cloud -- SAP AI Core -- SAP Generative AI Hub -- SAP Databricks -- SAP Datasphere -- SAP Analytics Cloud - -For detailed implementation guides and "how-to" tutorials, please refer to the official SAP documentation and related technical blog posts. \ No newline at end of file + +This reference architecture is realized through the following key SAP services and components: + +### Related Reference Architectures + +**SAP Business Data Cloud Series:** +- [Data Products in SAP Business Data Cloud](../1-data-products-in-sap-business-data-cloud/readme.md) - Understanding data products, their architecture, and consumption patterns +- [SAP Databricks in SAP BDC](../5-sap-databricks-in-business-data-cloud/readme.md) - Deep dive into SAP Databricks integration and use cases +- [Intelligent Applications in SAP Business Data Cloud](../2-intelligent-applications-by-sap/readme.md) - Pre-configured analytics and dashboards +- [Modernizing SAP BW with SAP Business Data Cloud](../4-modernizing-sap-bw-with-sap-bdc/readme.md) - Migration patterns and data product generation +- [Cloud Identity Services for BDC](../6-cloud-identity-services-bdc/readme.md) - Unified identity and access management + +**Generative AI and Machine Learning:** +- [Generative AI with SAP AI Core](../../RA0005/readme.md) - Comprehensive guide to GenAI patterns, RAG, and AI agents +- [Federated Machine Learning with SAP Datasphere](../../RA0003/readme.md) - ML integration across hyperscaler platforms + +### SAP Services and Documentation + +**SAP AI Foundation:** +- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) - Enterprise AI platform for model lifecycle management +- [SAP AI Launchpad](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-launchpad) - Multi-tenant SaaS for managing AI scenarios +- [Generative AI Hub in SAP AI Core](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/generative-ai-hub-in-sap-ai-core) - Access to foundation models and LLMs +- [SAP AI Core SDK](https://pypi.org/project/ai-core-sdk/) - Python SDK for programmatic AI Core integration + +**SAP Business Data Cloud:** +- [SAP Business Data Cloud Overview](https://www.sap.com/products/technology-platform/business-data-cloud.html) - Product overview and capabilities +- [SAP Datasphere](https://help.sap.com/docs/SAP_DATASPHERE) - Data management, modeling, and integration +- [SAP Analytics Cloud](https://www.sap.com/products/technology-platform/cloud-analytics.html) - Business intelligence and analytics +- [SAP Databricks Documentation](https://help.sap.com/docs/sap-datasphere/sap-datasphere-administration-guide-for-sap-datasphere/sap-databricks) - Integration guide for SAP Databricks + +**Development and Integration:** +- [SAP Cloud Application Programming Model (CAP)](https://cap.cloud.sap/docs/) - Framework for building enterprise applications +- [SAP AI SDK for JavaScript/TypeScript](https://github.com/SAP/ai-sdk-js) - SDK for integrating AI capabilities +- [SAP BDC Connect SDK](https://pypi.org/project/sap-bdc-connect-sdk/) - Python SDK for data product management + +### Learning Resources + +- [Introducing SAP Business Data Cloud](https://learning.sap.com/learning-journeys/introducing-sap-business-data-cloud) - Learning journey +- [SAP AI Core Tutorial](https://developers.sap.com/tutorials/ai-core-genaihub-provisioning.html) - Set up Generative AI Hub +- [SAP Community: Business Data Cloud](https://community.sap.com/topics/business-data-cloud) - Community discussions and blogs From 47c7f1de0496c2abf4333a9a523a38a80cb17d97 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Thu, 11 Dec 2025 15:13:38 -0800 Subject: [PATCH 12/19] Fix tags --- docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 1243f69b6c..25bcaa7168 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -17,14 +17,11 @@ keywords: - reference architecture sidebar_label: SAP Business Data Cloud powered by SAP AI Core tags: - - ai - data - - sap-ai-core - genai - databricks - - mlops - - architecture - - data-products + - bdc + - agents hide_table_of_contents: false hide_title: false toc_min_heading_level: 2 From b127e8988f23cd5e7e17fc681b75ddd0c8cfe847 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Tue, 20 Jan 2026 15:23:00 -0600 Subject: [PATCH 13/19] Reduce content and simplify BDC AI Core reference architecture - Removed text redundancies and reduced content size by ~30% - Removed tabs from the drawio diagram --- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 212 +++++------------- 1 file changed, 61 insertions(+), 151 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 25bcaa7168..61a7370a1a 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -4,12 +4,11 @@ slug: /ref-arch/f5b6b597a6/7 sidebar_position: 7 title: SAP Business Data Cloud powered by SAP AI Core description: >- - Strategic value and architectural patterns for integrating SAP Business Data Cloud with SAP AI Core and Generative AI Hub. Covers AI-Enhanced Data Products, model training in Databricks and serving in AI Core, batch and real-time consumption patterns, predictive insights, and autonomous process optimization with AI agents. + Architectural patterns for integrating SAP Business Data Cloud with SAP AI Core and Generative AI Hub. Covers AI-Enhanced Data Products, model training in Databricks and serving in AI Core, batch and real-time consumption patterns, and predictive insights. keywords: - sap business data cloud - sap ai core - generative ai hub - - tabular ai - sap databricks - data products - model lifecycle @@ -21,7 +20,6 @@ tags: - genai - databricks - bdc - - agents hide_table_of_contents: false hide_title: false toc_min_heading_level: 2 @@ -38,62 +36,55 @@ last_update: date: 2025-12-10 --- -Enterprises possess a wealth of invaluable business data within their SAP systems. However, activating this data for modern Artificial Intelligence is often a complex, disconnected, and risky endeavor. To stay competitive, organizations need a strategy to transform this data into reliable, governed, and actionable AI-driven insights that are deeply integrated with core business processes. +Enterprises possess a wealth of invaluable business data within their SAP systems. However, activating this data for modern Artificial Intelligence is often a complex, disconnected, and risky challenge. To stay competitive, organizations need a strategy to transform this data into reliable, governed, and actionable AI-driven insights that are deeply integrated with core business processes. -This reference architecture presents a cohesive vision for combining **SAP Business Data Cloud** with **SAP AI Foundation** (including SAP AI Core and the Generative AI Hub). The core architectural concept is the creation of **AI-Enhanced Data Products**—intelligent, context-aware, and dynamic assets that deliver trusted predictive insights and drive business automation at scale. This integrated approach enables the development of AI solutions that are **reliable, relevant, and responsible**, accelerating time-to-value and embedding intelligence directly into the enterprise. +This reference architecture presents a cohesive vision for combining **SAP Business Data Cloud** with **SAP AI Foundation** (including SAP AI Core and the Generative AI Hub). The core architectural concept is the creation of **AI-Enhanced Data Products**—intelligent, context-aware, and dynamic assets that deliver trusted predictive insights and drive business automation at scale. -**Future State: Real-Time Data Access via MCP** - -To further enhance real-time consumption patterns, SAP is developing a managed Model Context Protocol (MCP) component that will provide a standardized way to query derived data products directly from SAP Business Data Cloud. This future capability will simplify real-time data access for AI applications, enabling seamless integration of governed BDC data products into live business processes. +:::note Future State +SAP is developing a managed Model Context Protocol (MCP) component for standardized querying of BDC data products, simplifying real-time data access for AI applications. +::: ![drawio](drawio/bdc-ai-core-integration.drawio) ## The Architectural Blueprint -The architectural vision is centered on a powerful synergy: SAP Business Data Cloud provides the governed, business-ready data foundation, while SAP AI Foundation offers a comprehensive portfolio of enterprise-grade AI capabilities. This avoids a "one-size-fits-all" approach, recognizing that different business problems require different tools. The value lies in the comphreensive AI capabilities that allows data scientists and developers to move from data preparation to model deployment and inference. - -This blueprint is built on the principle of using each component for its primary strength: - -- **SAP Business Data Cloud (with SAP Databricks):** A comprehensive **enterprise data platform** that serves as the unified foundation for AI-driven business intelligence. It provides end-to-end capabilities for **discovering, connecting, preparing, exploring, curating, and governing** both SAP and non-SAP data sources through a semantically rich, business-context-aware layer. - -- **SAP AI Foundation (SAP AI Core & Generative AI Hub):** A comprehensive **enterprise AI platform** that offers a complete set of capabilities for managing the full lifecycle of machine learning and generative AI models. It supports scalable training, deployment, and monitoring, along with workflow orchestration, model versioning, and secure integration with SAP applications. SAP AI Core enables customers to use several models or allow customers to bring and operationalize their own AI models. These models can be accessed and governed through the Generative AI Hub, making them reusable services across the enterprise. +The architectural vision is centered on a powerful synergy: SAP Business Data Cloud provides the governed, business-ready data foundation, while SAP AI Foundation offers a comprehensive portfolio of enterprise-grade AI capabilities. The value lies in the comprehensive AI capabilities that allow data scientists and developers to move from data preparation to model deployment and inference. -## Design Principles and Unique Strengths of SAP AI Core +- **SAP Business Data Cloud (with SAP Databricks):** A comprehensive **enterprise data platform** serving as the unified foundation for AI-driven business intelligence. It provides end-to-end capabilities for **discovering, connecting, preparing, exploring, curating, and governing** both SAP and non-SAP data sources through a semantically rich, business-context-aware layer. -* **Rich Portfolio of GenAI Models:** Delivered via the Generative AI Hub and governed as reusable services. This provides a broader selection of foundation models tailored for SAP-centric workloads than is available on many general-purpose platforms. +- **SAP AI Foundation (SAP AI Core & Generative AI Hub):** A comprehensive **enterprise AI platform** for managing the full lifecycle of machine learning and generative AI models. It supports scalable training, deployment, monitoring, workflow orchestration, model versioning, and secure integration with SAP applications. -* **SAP-First Governance:** Model training, deployment, monitoring, and auditability are all aligned with SAP's identity, policy, and compliance expectations, ensuring enterprise-grade security. +### Design Principles of SAP AI Core -* **Operational Alignment:** Features such as deployment targets, access controls, observability, and cost management are specifically designed to integrate smoothly with existing SAP applications and business processes. - -* **Complementary by Design:** SAP AI Core is built to work alongside, not replace, existing data platforms. It complements lakehouse analytics and data science environments by operationalizing AI services *inside* the SAP landscape, which accelerates value realization. +- **Rich GenAI Portfolio:** Generative AI Hub provides a broad selection of foundation models governed as reusable services +- **SAP-First Governance:** Training, deployment, monitoring, and auditability aligned with SAP identity, policy, and compliance +- **Operational Alignment:** Deployment targets, access controls, observability, and cost management integrate with SAP applications +- **Complementary by Design:** Works alongside existing data platforms, operationalizing AI services inside the SAP landscape ## Key Architectural Patterns -To compete today, enterprises must activate their most valuable asset — their core business data in SAP — for modern AI. However, this is often a complex, disconnected, and risky endeavor. - -The strategy we're outlining provides a clear, governed path to solve this. It's built on the powerful synergy between SAP Business Data Cloud for our data foundation and SAP AI Foundation for our AI capabilities. +The following patterns provide a clear, governed path for activating SAP data for modern AI. -### Pattern 1: Train in Databricks, Serve in AI Core (The "Foundational Pattern") +### Pattern 1: Train in Databricks, Serve in AI Core **What:** A data scientist performs exploratory data science and model training in SAP Databricks, directly against governed, business-ready data products shared from SAP BDC Cockpit. During development, they engineer features within their notebooks—simple transformations stay with the model as preprocessing steps, while reusable features can be promoted back to BDC data products for broader consumption. The goal is to produce a high-quality, production-ready model. The key is a clean separation of roles: the data scientist creates the model or utilizes an existing model and releases the model artifacts to the unity catalog in Databricks, while a client-managed MLOps pipeline pulls these artifacts and handles validation and deployment to AI Core using SAP AI Core SDK. **Why:** This pattern provides separation of concerns—data scientists use a pro-code development environment to run experiments in Databricks while AI Core handles the production model serving. -### Pattern 2: The "Batch Consumption" Pattern (for Data Enrichment) +### Pattern 2: The Batch Consumption Pattern **What:** Building on Pattern 1's deployed model, this pattern enriches data products through high-throughput batch processing using scheduled or event-driven triggers (e.g., hourly ticket categorization, weekly sales forecasts). The model reads from a source dataset and writes predictions back—either as new columns in the existing dataset or as a separate data product in BDC. -**Trigger Options:** The automated pipeline can decide when batch processes run, providing full control and flexibility for integration: -* **Recurring:** Schedule periodic batch jobs to continuously process new or updated records as they arrive. BDC's Delta Lake foundation enables efficient incremental processing through Change Data Feed (CDF), ensuring only new data is processed rather than rescanning entire datasets. -* **One-Time:** Trigger bulk processing for historical data or initial data product creation across entire datasets. -* **Event-Driven:** Build logic to automatically trigger re-scoring in response to events, such as the deployment of a new model (via Pattern 1), ensuring insights reflect the latest AI logic. +**Trigger Options:** +* **Recurring:** Schedule periodic batch jobs with efficient incremental processing through Delta Lake's Change Data Feed (CDF) +* **One-Time:** Trigger bulk processing for historical data or initial data product creation +* **Event-Driven:** Automatically trigger re-scoring in response to events such as new model deployment -**Workflow Implementation:** Use **SAP AI Core workflow executions** for batch processing—distinct from real-time endpoints and optimized for high-throughput operations. Your **containerized workflow** reads from configured object stores (S3, Azure Blob), processes data, and writes predictions back. Schedule jobs with **cron in SAP AI Launchpad** for simple recurring tasks, or use the **SAP AI Core SDK** for client-managed programmatic control and integration with existing pipelines. +**Workflow Implementation:** Use **SAP AI Core workflow executions** for batch processing—distinct from real-time endpoints and optimized for high-throughput operations. Schedule jobs with **cron in SAP AI Launchpad** for simple recurring tasks, or use the **SAP AI Core SDK** for programmatic control. -**Why:** This architecture provides a robust, governable, and efficient way to handle all non-real-time AI processing. It uses the *same, single, governed model* from AI Core for all batch scenarios, ensuring that whether you are processing a small batch of new records or re-scoring an entire table, you are using the same "reliable" and "responsible" AI logic. It separates high-throughput batch workloads from low-latency online serving, ensuring the right resources are used for the right job, promoting both performance and cost-efficiency. +**Why:** This architecture provides a robust, governable way to handle non-real-time AI processing. It uses the *same governed model* from AI Core for all batch scenarios, separating high-throughput batch workloads from low-latency online serving. -### Pattern 3: The "Real-Time Consumption" Pattern (for Embedded Intelligence) +### Pattern 3: The Real-Time Consumption Pattern **What:** Building on Pattern 1's deployed model, this pattern embeds AI into live business processes for immediate predictions. Applications make low-latency API calls to the **deployment endpoint** (distinct from Pattern 2's batch *executions*), receiving instant insights to drive decisions—like checking risk scores before posting a sales order or providing recommendations as a page loads. @@ -105,154 +96,73 @@ The strategy we're outlining provides a clear, governed path to solve this. It's **Implementation with SAP Cloud Application Programming Model (CAP):** -CAP provides a natural fit for implementing Pattern 3, offering significant advantages for deployments: +CAP provides a natural fit for implementing Pattern 3: -* **Integrated Data Access:** CAP applications can seamlessly query SAP Datasphere (which federates BDC data products) and combine this with real-time AI predictions in a single request-response cycle, eliminating the need for separate data and inference layers. In the future, the SAP-managed MCP component will further streamline data retrieval by providing a standardized query interface for BDC data products. Applications can then combine this contextual business data with AI Core predictions (via separate API calls) to create a complete real-time intelligence flow. -* **Built-in Governance:** Authorization, authentication, and audit logging align automatically with SAP standards—the same security model protects both your data retrieval and AI inference calls. -* **Simplified Development:** Developers work within familiar SAP frameworks (CDS models, OData services) rather than managing low-level HTTP clients, reducing integration complexity and accelerating time-to-market. -* **Enterprise-Ready:** CAP applications deploy naturally into SAP BTP with built-in observability, scaling, and operational tooling—no additional infrastructure setup required. +* **Integrated Data Access:** CAP applications can seamlessly query SAP Datasphere (which federates BDC data products) and combine this with real-time AI predictions in a single request-response cycle +* **Built-in Governance:** Authorization, authentication, and audit logging align automatically with SAP standards +* **Simplified Development:** Developers work within familiar SAP frameworks (CDS models, OData services) rather than managing low-level HTTP clients +* **Enterprise-Ready:** CAP applications deploy naturally into SAP BTP with built-in observability, scaling, and operational tooling -## Business Problem 1: AI-Enhanced Predictive Insights +## Business Problem: AI-Enhanced Predictive Insights -To make these patterns concrete, let's walk through a tangible, high-value example: **Improving Cash Flow with AI-Enhanced Payment Delay Predictions.** +To make these patterns concrete, let's walk through a tangible example: **Improving Cash Flow with AI-Enhanced Payment Delay Predictions.** -A large enterprise's finance department struggles with reactive cash flow management. They can see which payments are overdue, but they lack the foresight to act proactively. They need to not only *predict* which payments are likely to be delayed but also *understand why* so the collections team can prioritize their efforts and engage with customers in a more informed and targeted manner. +A large enterprise's finance department struggles with reactive cash flow management. They can see which payments are overdue, but they lack the foresight to act proactively. They need to not only *predict* which payments are likely to be delayed but also *understand why* so the collections team can prioritize their efforts. -### The Solution: Building an AI-Enhanced Data Product - -An AI-Enhanced Data Product is created by following the defined architectural patterns, with clear roles for each persona. +### Solution: Building an AI-Enhanced Data Product **1. Model Development (Persona: Data Scientist in SAP BDC)** * The data scientist explores data in the SAP BDC catalog, identifying the `Entry View Journal Entry` data product. -* Working in an SAP Databricks notebook, they realize a simple prediction (the "what") is insufficient. They must also productionalize the explanation (the "why"). -* They prototype an **end-to-end prediction and explanation pipeline**. This pipeline: - 1. Trains an **XGBoost** model to get the prediction. - 2. Uses the **SHAP** library to calculate feature importance. - 3. Calls the **Generative AI Hub** (via the SAP AI SDK) to translate the SHAP values into a human-readable explanation (e.g., *"This payment is predicted to be 15 days late, primarily due to past payment behavior..."*). -* This entire, self-contained pipeline is saved as a single deployable asset. This completes the "Data Science on BDC" (development) part of **Pattern 1**. +* Working in an SAP Databricks notebook, they prototype an **end-to-end prediction and explanation pipeline**: + 1. Trains an **XGBoost** model to get the prediction + 2. Uses the **SHAP** library to calculate feature importance + 3. Calls the **Generative AI Hub** (via the SAP AI SDK) to translate the SHAP values into a human-readable explanation +* This pipeline is saved as a single deployable asset, completing the development part of **Pattern 1**. **2. Enterprise Deployment (Personas: ML Engineer & Data Scientist)** -* This step follows the "MLOps on AI Core" (deployment) part of **Pattern 1: Train in Databricks, Serve in AI Core**. -* The **ML Engineer** *enables* this by first building a reusable serving template in **SAP AI Core** capable of running this complex (XGBoost + SHAP + GenAI) pipeline. -* The **Data Scientist**, still in their BDC notebook, uses the **SAP AI Core SDK** to trigger the deployment. They pass their saved pipeline asset to the ML Engineer's template. -* **SAP AI Core** automatically builds, containerizes, and deploys this entire pipeline, exposing a single, scalable, and governed API endpoint. This endpoint, when called, now returns the *full* payload: both the prediction and the explanation. - -**3. Operationalizing the "Why" (Personas: App Developer, Business User)** - -The deployed API is now operationalized using *both* consumption patterns to solve the business problem: - -* **Pattern 2: The Batch Consumption Pattern:** - * A **recurring workflow** is scheduled in **SAP AI Launchpad**. It follows this pattern to execute the native batch pipeline, reading all new invoices. - * It creates the **"Enriched Payment Forecasts" data product** in BDC, which now includes both the risk score *and* the human-readable explanation. - * The **Business User** opens their **SAP Analytics Cloud** dashboard, which reads this data product to see a fully-explained, prioritized worklist. - -* **Pattern 3: The Real-Time Consumption Pattern:** - * An **Application Developer** builds a Fiori app for the collections team. - * When a user opens a customer account, the app makes a *live call* to the *same* AI Core API for that customer's outstanding invoices. - * This pattern provides the team with instant, on-demand predictions and explanations to guide their conversation. - -### The Value Proposition (The "Why") - -This end-to-end scenario delivers value at multiple levels: - -- **For the Business:** The finance team moves from reactive to proactive, improving cash flow and enabling more meaningful customer interactions. The business gains a trusted, explainable AI solution, not a "black box." - -- **For the Data Scientist:** They can innovate rapidly in the agile Databricks environment while leveraging powerful, enterprise-grade AI capabilities from SAP AI Core without needing to be an expert in Kubernetes or API management. +* The **ML Engineer** builds a reusable serving template in **SAP AI Core** capable of running this complex (XGBoost + SHAP + GenAI) pipeline. +* The **Data Scientist** uses the **SAP AI Core SDK** to trigger the deployment, passing their saved pipeline asset to the ML Engineer's template. +* **SAP AI Core** automatically builds, containerizes, and deploys this pipeline, exposing a single, scalable, governed API endpoint that returns both prediction and explanation. -- **For IT & Governance:** The entire process is governed. Data access is controlled, the model is monitored, and the resulting data product is a managed asset. The architecture provides the robustness and auditability required for a mission-critical financial process. +**3. Operationalizing the Insights** -## Business Problem 2: Autonomous Process Optimization with AI Agents +The deployed API is operationalized using *both* consumption patterns: -Beyond providing insights for users to act on, the architecture's true power is realized when the foundational patterns are combined to create **specialized AI data agents**. These are autonomous constructs that monitor, analyze, and act on business events with minimal human intervention, moving the business from a reactive to a proactive and automated posture. +* **Pattern 2 (Batch):** A recurring workflow scheduled in SAP AI Launchpad creates the **"Enriched Payment Forecasts" data product** in BDC, which business users view in their SAP Analytics Cloud dashboard. -A practical example is a **"Cash Flow Optimization Agent"**, which builds directly on the previous scenario: +* **Pattern 3 (Real-Time):** A Fiori app for the collections team makes live calls to the same AI Core API when a user opens a customer account, providing instant predictions. -1. **Monitor (Powered by Pattern 2):** The agent continuously observes the "Enriched Payment Forecasts" data product. This data is updated daily via **Pattern 2: The Batch Consumption Pattern**. +### Value Delivered -2. **Analyze (Powered by Pattern 3):** When a high-risk invoice is detected, the agent *executes* **Pattern 3: The Real-Time Consumption Pattern**. It makes a real-time call to the deployed API (the one created by Pattern 1) to get an immediate, deep explanation. - -3. **Act (The Next Step):** Based on the analysis from Pattern 3, the agent triggers a workflow in SAP BTP to assign a task or uses a generative AI model to draft a personalized outreach email. - -### Key Recommendations and Best Practices for Batch Inference - -* **Utilize Workflow Executions:** Orchestrate batch inference as pipelines/workflow executions in SAP AI Core to manage high-volume, offline processing efficiently. -* **Scheduling:** Schedule recurring batch jobs using cron specifications in SAP AI Launchpad to automate periodic tasks (e.g., weekly or monthly reports). -* **Efficiency:** Prioritize throughput over latency; leverage GPUs and other compute resources efficiently for large datasets. -* **Data Handling:** Design batch programs to process data in bulk, reading from and writing to configured object stores (e.g., S3). Produce an output dataset of predictions for downstream consumption or as a governed data product. -* **Containerization:** Package batch inference logic in a user-supplied Docker image containing the necessary code and dependencies (e.g., PyTorch, TensorFlow) for execution in SAP AI Core. -* **API-First Approach:** Manage and trigger configurations and executions via SAP AI Core APIs or the Python SDK to integrate with external applications and pipelines. -* **Separate from Online Serving:** Do not use real-time serving endpoints for batch; online endpoints are optimized for low-latency single or small-batch requests, whereas workflow executions are designed for high-volume, offline processing. -* **Resource Management:** Leverage platform capabilities to dynamically scale compute resources based on job demands to ensure cost-efficiency and performance. - -## Summary of Roles and Responsibilities - -- **Data Scientists:** Primarily work within **SAP Business Data Cloud (Databricks)** for data exploration, feature engineering, and rapid model experimentation. -- **ML Engineers & IT Operations:** Primarily work with **SAP AI Core** to manage production model deployments, monitor performance, ensure governance, and maintain the operational integrity of AI services. -- **Application Developers & Business Users:** Consume the final AI-Enhanced Data Products and AI-powered applications through various channels, including SAP Analytics Cloud, custom BTP applications, or integrated line-of-business solutions. - -## Platform Selection Guide - -### Use **Databricks** for: -- **Data science experimentation** and rapid prototyping -- **Quick validation** by data scientist personas -- Integrated data science workflows with immediate feedback - -### Use **AI Core** for: -- **Production-ready models** requiring enterprise-grade serving -- Access to **broader LLM ecosystem** and **SAP's RPT-1** foundational model -- Scenarios requiring **production infrastructure** and planned **BDC integrations** - -## Key Differentiators - -| Aspect | AI Core | Databricks | -|--------|---------|------------| -| **Use Case** | Production deployment | Experimentation & prototyping | -| **Infrastructure** | Enterprise-grade serving | Integrated development environment | -| **Model Access** | Broad LLM ecosystem + RPT-1 | Selected LLMs | -| **Speed to Value** | Production-ready deployment | Rapid prototyping | -| **Integration** | Planned BDC integrations | Native data science workflows | +- **For the Business:** The finance team moves from reactive to proactive, improving cash flow with trusted, explainable AI +- **For the Data Scientist:** Rapid innovation in Databricks while leveraging enterprise-grade AI capabilities from SAP AI Core +- **For IT & Governance:** Controlled data access, monitored models, and managed data products with enterprise auditability ## Components and Further Reading - -This reference architecture is realized through the following key SAP services and components: - ### Related Reference Architectures **SAP Business Data Cloud Series:** -- [Data Products in SAP Business Data Cloud](../1-data-products-in-sap-business-data-cloud/readme.md) - Understanding data products, their architecture, and consumption patterns -- [SAP Databricks in SAP BDC](../5-sap-databricks-in-business-data-cloud/readme.md) - Deep dive into SAP Databricks integration and use cases -- [Intelligent Applications in SAP Business Data Cloud](../2-intelligent-applications-by-sap/readme.md) - Pre-configured analytics and dashboards -- [Modernizing SAP BW with SAP Business Data Cloud](../4-modernizing-sap-bw-with-sap-bdc/readme.md) - Migration patterns and data product generation -- [Cloud Identity Services for BDC](../6-cloud-identity-services-bdc/readme.md) - Unified identity and access management +- [Data Products in SAP Business Data Cloud](../1-data-products-in-sap-business-data-cloud/readme.md) +- [SAP Databricks in SAP BDC](../5-sap-databricks-in-business-data-cloud/readme.md) +- [Intelligent Applications in SAP Business Data Cloud](../2-intelligent-applications-by-sap/readme.md) +- [Modernizing SAP BW with SAP Business Data Cloud](../4-modernizing-sap-bw-with-sap-bdc/readme.md) +- [Cloud Identity Services for BDC](../6-cloud-identity-services-bdc/readme.md) **Generative AI and Machine Learning:** -- [Generative AI with SAP AI Core](../../RA0005/readme.md) - Comprehensive guide to GenAI patterns, RAG, and AI agents -- [Federated Machine Learning with SAP Datasphere](../../RA0003/readme.md) - ML integration across hyperscaler platforms - -### SAP Services and Documentation - -**SAP AI Foundation:** -- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) - Enterprise AI platform for model lifecycle management -- [SAP AI Launchpad](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-launchpad) - Multi-tenant SaaS for managing AI scenarios -- [Generative AI Hub in SAP AI Core](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/generative-ai-hub-in-sap-ai-core) - Access to foundation models and LLMs -- [SAP AI Core SDK](https://pypi.org/project/ai-core-sdk/) - Python SDK for programmatic AI Core integration +- [Generative AI with SAP AI Core](../../RA0005/readme.md) - GenAI patterns, RAG, and AI agents +- [Federated Machine Learning with SAP Datasphere](../../RA0003/readme.md) -**SAP Business Data Cloud:** -- [SAP Business Data Cloud Overview](https://www.sap.com/products/technology-platform/business-data-cloud.html) - Product overview and capabilities -- [SAP Datasphere](https://help.sap.com/docs/SAP_DATASPHERE) - Data management, modeling, and integration -- [SAP Analytics Cloud](https://www.sap.com/products/technology-platform/cloud-analytics.html) - Business intelligence and analytics -- [SAP Databricks Documentation](https://help.sap.com/docs/sap-datasphere/sap-datasphere-administration-guide-for-sap-datasphere/sap-databricks) - Integration guide for SAP Databricks +### SAP Services -**Development and Integration:** -- [SAP Cloud Application Programming Model (CAP)](https://cap.cloud.sap/docs/) - Framework for building enterprise applications -- [SAP AI SDK for JavaScript/TypeScript](https://github.com/SAP/ai-sdk-js) - SDK for integrating AI capabilities -- [SAP BDC Connect SDK](https://pypi.org/project/sap-bdc-connect-sdk/) - Python SDK for data product management +- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) | [SAP AI Launchpad](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-launchpad) | [Generative AI Hub](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/generative-ai-hub-in-sap-ai-core) +- [SAP Business Data Cloud](https://www.sap.com/products/technology-platform/business-data-cloud.html) | [SAP Datasphere](https://help.sap.com/docs/SAP_DATASPHERE) | [SAP Databricks](https://help.sap.com/docs/sap-datasphere/sap-datasphere-administration-guide-for-sap-datasphere/sap-databricks) +- [SAP AI Core SDK (Python)](https://pypi.org/project/ai-core-sdk/) | [SAP AI SDK (JS/TS)](https://github.com/SAP/ai-sdk-js) | [SAP BDC Connect SDK](https://pypi.org/project/sap-bdc-connect-sdk/) +- [SAP CAP](https://cap.cloud.sap/docs/) | [SAP Analytics Cloud](https://www.sap.com/products/technology-platform/cloud-analytics.html) ### Learning Resources - [Introducing SAP Business Data Cloud](https://learning.sap.com/learning-journeys/introducing-sap-business-data-cloud) - Learning journey - [SAP AI Core Tutorial](https://developers.sap.com/tutorials/ai-core-genaihub-provisioning.html) - Set up Generative AI Hub -- [SAP Community: Business Data Cloud](https://community.sap.com/topics/business-data-cloud) - Community discussions and blogs From 1375ca9fda0820c1eae3ea57e3c4b451d0b01c4f Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Wed, 21 Jan 2026 13:53:25 -0600 Subject: [PATCH 14/19] Streamline BDC powered by AI Core reference architecture - Remove 6 unreferenced links from Components section (Intelligent Apps, BW Modernization, Cloud Identity, Federated ML, BDC Connect SDK, SAP Community) - Consolidate SAP AI Foundation description with Design Principles into a single concise paragraph - Remove redundant Best Practices for Batch Inference section (content already covered in Pattern 2) --- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 65 +------------------ 1 file changed, 3 insertions(+), 62 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 25bcaa7168..6b789e72e9 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -56,23 +56,11 @@ This blueprint is built on the principle of using each component for its primary - **SAP Business Data Cloud (with SAP Databricks):** A comprehensive **enterprise data platform** that serves as the unified foundation for AI-driven business intelligence. It provides end-to-end capabilities for **discovering, connecting, preparing, exploring, curating, and governing** both SAP and non-SAP data sources through a semantically rich, business-context-aware layer. -- **SAP AI Foundation (SAP AI Core & Generative AI Hub):** A comprehensive **enterprise AI platform** that offers a complete set of capabilities for managing the full lifecycle of machine learning and generative AI models. It supports scalable training, deployment, and monitoring, along with workflow orchestration, model versioning, and secure integration with SAP applications. SAP AI Core enables customers to use several models or allow customers to bring and operationalize their own AI models. These models can be accessed and governed through the Generative AI Hub, making them reusable services across the enterprise. - -## Design Principles and Unique Strengths of SAP AI Core - -* **Rich Portfolio of GenAI Models:** Delivered via the Generative AI Hub and governed as reusable services. This provides a broader selection of foundation models tailored for SAP-centric workloads than is available on many general-purpose platforms. - -* **SAP-First Governance:** Model training, deployment, monitoring, and auditability are all aligned with SAP's identity, policy, and compliance expectations, ensuring enterprise-grade security. - -* **Operational Alignment:** Features such as deployment targets, access controls, observability, and cost management are specifically designed to integrate smoothly with existing SAP applications and business processes. - -* **Complementary by Design:** SAP AI Core is built to work alongside, not replace, existing data platforms. It complements lakehouse analytics and data science environments by operationalizing AI services *inside* the SAP landscape, which accelerates value realization. +- **SAP AI Foundation (SAP AI Core & Generative AI Hub):** An **enterprise AI platform** for managing the full lifecycle of ML and GenAI models—training, deployment, monitoring, and governance. Customers can use SAP-provided models or bring their own, with the Generative AI Hub providing a rich portfolio of foundation models as reusable, governed services. SAP AI Core is designed to complement (not replace) existing data platforms, integrating seamlessly with SAP identity, policy, and compliance frameworks while providing enterprise-grade access controls, observability, and cost management. ## Key Architectural Patterns -To compete today, enterprises must activate their most valuable asset — their core business data in SAP — for modern AI. However, this is often a complex, disconnected, and risky endeavor. - -The strategy we're outlining provides a clear, governed path to solve this. It's built on the powerful synergy between SAP Business Data Cloud for our data foundation and SAP AI Foundation for our AI capabilities. +The strategy outlined in this reference architecture provides a clear, governed path to activate your most valuable asset—core business data in SAP—for modern AI. It's built on the powerful synergy between SAP Business Data Cloud for our data foundation and SAP AI Foundation for our AI capabilities. ### Pattern 1: Train in Databricks, Serve in AI Core (The "Foundational Pattern") @@ -112,7 +100,7 @@ CAP provides a natural fit for implementing Pattern 3, offering significant adva * **Simplified Development:** Developers work within familiar SAP frameworks (CDS models, OData services) rather than managing low-level HTTP clients, reducing integration complexity and accelerating time-to-market. * **Enterprise-Ready:** CAP applications deploy naturally into SAP BTP with built-in observability, scaling, and operational tooling—no additional infrastructure setup required. -## Business Problem 1: AI-Enhanced Predictive Insights +## Business Problem: AI-Enhanced Predictive Insights To make these patterns concrete, let's walk through a tangible, high-value example: **Improving Cash Flow with AI-Enhanced Payment Delay Predictions.** @@ -163,47 +151,6 @@ This end-to-end scenario delivers value at multiple levels: - **For IT & Governance:** The entire process is governed. Data access is controlled, the model is monitored, and the resulting data product is a managed asset. The architecture provides the robustness and auditability required for a mission-critical financial process. -## Business Problem 2: Autonomous Process Optimization with AI Agents - -Beyond providing insights for users to act on, the architecture's true power is realized when the foundational patterns are combined to create **specialized AI data agents**. These are autonomous constructs that monitor, analyze, and act on business events with minimal human intervention, moving the business from a reactive to a proactive and automated posture. - -A practical example is a **"Cash Flow Optimization Agent"**, which builds directly on the previous scenario: - -1. **Monitor (Powered by Pattern 2):** The agent continuously observes the "Enriched Payment Forecasts" data product. This data is updated daily via **Pattern 2: The Batch Consumption Pattern**. - -2. **Analyze (Powered by Pattern 3):** When a high-risk invoice is detected, the agent *executes* **Pattern 3: The Real-Time Consumption Pattern**. It makes a real-time call to the deployed API (the one created by Pattern 1) to get an immediate, deep explanation. - -3. **Act (The Next Step):** Based on the analysis from Pattern 3, the agent triggers a workflow in SAP BTP to assign a task or uses a generative AI model to draft a personalized outreach email. - -### Key Recommendations and Best Practices for Batch Inference - -* **Utilize Workflow Executions:** Orchestrate batch inference as pipelines/workflow executions in SAP AI Core to manage high-volume, offline processing efficiently. -* **Scheduling:** Schedule recurring batch jobs using cron specifications in SAP AI Launchpad to automate periodic tasks (e.g., weekly or monthly reports). -* **Efficiency:** Prioritize throughput over latency; leverage GPUs and other compute resources efficiently for large datasets. -* **Data Handling:** Design batch programs to process data in bulk, reading from and writing to configured object stores (e.g., S3). Produce an output dataset of predictions for downstream consumption or as a governed data product. -* **Containerization:** Package batch inference logic in a user-supplied Docker image containing the necessary code and dependencies (e.g., PyTorch, TensorFlow) for execution in SAP AI Core. -* **API-First Approach:** Manage and trigger configurations and executions via SAP AI Core APIs or the Python SDK to integrate with external applications and pipelines. -* **Separate from Online Serving:** Do not use real-time serving endpoints for batch; online endpoints are optimized for low-latency single or small-batch requests, whereas workflow executions are designed for high-volume, offline processing. -* **Resource Management:** Leverage platform capabilities to dynamically scale compute resources based on job demands to ensure cost-efficiency and performance. - -## Summary of Roles and Responsibilities - -- **Data Scientists:** Primarily work within **SAP Business Data Cloud (Databricks)** for data exploration, feature engineering, and rapid model experimentation. -- **ML Engineers & IT Operations:** Primarily work with **SAP AI Core** to manage production model deployments, monitor performance, ensure governance, and maintain the operational integrity of AI services. -- **Application Developers & Business Users:** Consume the final AI-Enhanced Data Products and AI-powered applications through various channels, including SAP Analytics Cloud, custom BTP applications, or integrated line-of-business solutions. - -## Platform Selection Guide - -### Use **Databricks** for: -- **Data science experimentation** and rapid prototyping -- **Quick validation** by data scientist personas -- Integrated data science workflows with immediate feedback - -### Use **AI Core** for: -- **Production-ready models** requiring enterprise-grade serving -- Access to **broader LLM ecosystem** and **SAP's RPT-1** foundational model -- Scenarios requiring **production infrastructure** and planned **BDC integrations** - ## Key Differentiators | Aspect | AI Core | Databricks | @@ -224,13 +171,9 @@ This reference architecture is realized through the following key SAP services a **SAP Business Data Cloud Series:** - [Data Products in SAP Business Data Cloud](../1-data-products-in-sap-business-data-cloud/readme.md) - Understanding data products, their architecture, and consumption patterns - [SAP Databricks in SAP BDC](../5-sap-databricks-in-business-data-cloud/readme.md) - Deep dive into SAP Databricks integration and use cases -- [Intelligent Applications in SAP Business Data Cloud](../2-intelligent-applications-by-sap/readme.md) - Pre-configured analytics and dashboards -- [Modernizing SAP BW with SAP Business Data Cloud](../4-modernizing-sap-bw-with-sap-bdc/readme.md) - Migration patterns and data product generation -- [Cloud Identity Services for BDC](../6-cloud-identity-services-bdc/readme.md) - Unified identity and access management **Generative AI and Machine Learning:** - [Generative AI with SAP AI Core](../../RA0005/readme.md) - Comprehensive guide to GenAI patterns, RAG, and AI agents -- [Federated Machine Learning with SAP Datasphere](../../RA0003/readme.md) - ML integration across hyperscaler platforms ### SAP Services and Documentation @@ -249,10 +192,8 @@ This reference architecture is realized through the following key SAP services a **Development and Integration:** - [SAP Cloud Application Programming Model (CAP)](https://cap.cloud.sap/docs/) - Framework for building enterprise applications - [SAP AI SDK for JavaScript/TypeScript](https://github.com/SAP/ai-sdk-js) - SDK for integrating AI capabilities -- [SAP BDC Connect SDK](https://pypi.org/project/sap-bdc-connect-sdk/) - Python SDK for data product management ### Learning Resources - [Introducing SAP Business Data Cloud](https://learning.sap.com/learning-journeys/introducing-sap-business-data-cloud) - Learning journey - [SAP AI Core Tutorial](https://developers.sap.com/tutorials/ai-core-genaihub-provisioning.html) - Set up Generative AI Hub -- [SAP Community: Business Data Cloud](https://community.sap.com/topics/business-data-cloud) - Community discussions and blogs From a3a517aaa3c783d140ac91de0b9d58f9ae86a040 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Wed, 21 Jan 2026 14:03:05 -0600 Subject: [PATCH 15/19] Update BDC AI Core integration diagram --- .../drawio/bdc-ai-core-integration.drawio | 4810 ++--------------- 1 file changed, 323 insertions(+), 4487 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio index ba1b001843..3b133463df 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio @@ -1,4487 +1,323 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + \ No newline at end of file From 29b401cffa250e89fa65f26385e843c669594835 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Mon, 26 Jan 2026 11:35:54 -0600 Subject: [PATCH 16/19] Change from "CAP App" to "BTP Extension applications" to make it more inclusive of other application types like Node.js, Java, Python apps/services. --- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 13 +++---------- 1 file changed, 3 insertions(+), 10 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 61a7370a1a..7daff590db 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -40,10 +40,6 @@ Enterprises possess a wealth of invaluable business data within their SAP system This reference architecture presents a cohesive vision for combining **SAP Business Data Cloud** with **SAP AI Foundation** (including SAP AI Core and the Generative AI Hub). The core architectural concept is the creation of **AI-Enhanced Data Products**—intelligent, context-aware, and dynamic assets that deliver trusted predictive insights and drive business automation at scale. -:::note Future State -SAP is developing a managed Model Context Protocol (MCP) component for standardized querying of BDC data products, simplifying real-time data access for AI applications. -::: - ![drawio](drawio/bdc-ai-core-integration.drawio) ## The Architectural Blueprint @@ -94,14 +90,11 @@ The following patterns provide a clear, governed path for activating SAP data fo **Why:** This pattern creates a **reusable AI asset**—the single model from Pattern 1 serves both massive batch jobs (Pattern 2) and critical real-time processes, embedding intelligence directly into enterprise operations without duplication. -**Implementation with SAP Cloud Application Programming Model (CAP):** +**Implementation with SAP BTP Extension Applications:** -CAP provides a natural fit for implementing Pattern 3: +SAP BTP extension applications—built with CAP, Node.js, Java, Python, or other runtimes—are well suited for this pattern. The BTP destination service manages connectivity to AI Core, allowing applications to call the deployment endpoint directly. -* **Integrated Data Access:** CAP applications can seamlessly query SAP Datasphere (which federates BDC data products) and combine this with real-time AI predictions in a single request-response cycle -* **Built-in Governance:** Authorization, authentication, and audit logging align automatically with SAP standards -* **Simplified Development:** Developers work within familiar SAP frameworks (CDS models, OData services) rather than managing low-level HTTP clients -* **Enterprise-Ready:** CAP applications deploy naturally into SAP BTP with built-in observability, scaling, and operational tooling +This approach enables applications to query SAP Datasphere for BDC data products and obtain AI predictions within the same request cycle. Authorization, authentication, and audit logging align with BTP security standards, while deployment benefits from BTP's built-in observability and scaling capabilities. The pattern also supports AI agents that require real-time model inference combined with SAP data access. ## Business Problem: AI-Enhanced Predictive Insights From 5855c6088f693d0944aae8ddc6888f455d644609 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Wed, 21 Jan 2026 14:53:18 -0600 Subject: [PATCH 17/19] fix: correct links at bottom of readme.md --- .../RA0013/7-bdc-powered-by-ai-core/readme.md | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 6b789e72e9..8628e0019c 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -178,22 +178,17 @@ This reference architecture is realized through the following key SAP services a ### SAP Services and Documentation **SAP AI Foundation:** -- [SAP AI Core](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core) - Enterprise AI platform for model lifecycle management -- [SAP AI Launchpad](https://discovery-center.cloud.sap/serviceCatalog/sap-ai-launchpad) - Multi-tenant SaaS for managing AI scenarios +- [SAP AI Core](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/what-is-sap-ai-core) - Enterprise AI platform for model lifecycle management +- [SAP AI Launchpad](https://help.sap.com/docs/ai-launchpad/sap-ai-launchpad/sap-ai-launchpad-overview?q=SAP+AI+Launchpad) - Multi-tenant SaaS for managing AI scenarios - [Generative AI Hub in SAP AI Core](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/generative-ai-hub-in-sap-ai-core) - Access to foundation models and LLMs -- [SAP AI Core SDK](https://pypi.org/project/ai-core-sdk/) - Python SDK for programmatic AI Core integration +- [SAP AI Core SDK](https://help.sap.com/doc/generative-ai-hub-sdk/CLOUD/en-US/_reference/README_sphynx.html) - Python SDK for programmatic AI Core integration **SAP Business Data Cloud:** -- [SAP Business Data Cloud Overview](https://www.sap.com/products/technology-platform/business-data-cloud.html) - Product overview and capabilities +- [SAP Business Data Cloud Overview](https://www.sap.com/products/data-cloud.html) - Product overview and capabilities - [SAP Datasphere](https://help.sap.com/docs/SAP_DATASPHERE) - Data management, modeling, and integration - [SAP Analytics Cloud](https://www.sap.com/products/technology-platform/cloud-analytics.html) - Business intelligence and analytics -- [SAP Databricks Documentation](https://help.sap.com/docs/sap-datasphere/sap-datasphere-administration-guide-for-sap-datasphere/sap-databricks) - Integration guide for SAP Databricks +- [SAP Databricks Documentation](http://help.sap.com/docs/business-data-cloud/sap-databricks/introducing-sap-databricks) - Integration guide for SAP Databricks **Development and Integration:** - [SAP Cloud Application Programming Model (CAP)](https://cap.cloud.sap/docs/) - Framework for building enterprise applications - [SAP AI SDK for JavaScript/TypeScript](https://github.com/SAP/ai-sdk-js) - SDK for integrating AI capabilities - -### Learning Resources - -- [Introducing SAP Business Data Cloud](https://learning.sap.com/learning-journeys/introducing-sap-business-data-cloud) - Learning journey -- [SAP AI Core Tutorial](https://developers.sap.com/tutorials/ai-core-genaihub-provisioning.html) - Set up Generative AI Hub From 0d76c1c87fce3df05e35ec2af2f6aeb85c94d058 Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Tue, 27 Jan 2026 17:04:02 -0600 Subject: [PATCH 18/19] Improvements --- .../drawio/bdc-ai-core-integration.drawio | 413 +++++++++--------- 1 file changed, 209 insertions(+), 204 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio index 3b133463df..bb2acc5db0 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/drawio/bdc-ai-core-integration.drawio @@ -1,32 +1,125 @@ - + - - + + - - + + - - + + - - + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + - - + + - - + + + + + + + + + + + + + + + + + + + + + + + + + - - + + - - + + + + + @@ -40,10 +133,10 @@ - - + + - + @@ -58,23 +151,13 @@ - - - - - - - + + - - - - + + - - - - + @@ -82,17 +165,17 @@ - - + + - - + + - - + + - - + + @@ -100,20 +183,20 @@ - + - + - + - - + + - - + + @@ -121,201 +204,123 @@ - + - + - - - - - - - - - - - - - - - - - - - - - - - - - + + + + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + - + + - - + + - - + + - - + + - - + + - - - - - - - - - - - + + - + - + + - - + + + + + - + - - - - - + + - - + + - - + + - - + + - - + + - - + + - - + + - - + + - - + + - - + + - - + + - - + + - + - - - - - + - - - - + + + + - + - + + + + + + - - + + + + + + + + + + + From 3795bab7319f70f70dc89f49edfefee9e93dd9cb Mon Sep 17 00:00:00 2001 From: Guilherme Segantini Date: Wed, 28 Jan 2026 12:26:59 -0600 Subject: [PATCH 19/19] Last date updated --- docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md index 7daff590db..f8a639394d 100644 --- a/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md +++ b/docs/ref-arch/RA0013/7-bdc-powered-by-ai-core/readme.md @@ -32,8 +32,8 @@ contributors: - jmsrpp - anbazhagan-uma last_update: - author: seeobjectively - date: 2025-12-10 + author: guilherme-segantini + date: 2026-01-28 --- Enterprises possess a wealth of invaluable business data within their SAP systems. However, activating this data for modern Artificial Intelligence is often a complex, disconnected, and risky challenge. To stay competitive, organizations need a strategy to transform this data into reliable, governed, and actionable AI-driven insights that are deeply integrated with core business processes.