About Modernizing SAP With Oracle AI Data Platform and Oracle Fusion Cloud Applications

Many enterprises run SAP (ECC or S/4HANA) as the operational system-of-record while adopting Oracle technologies - most commonly Oracle Fusion Cloud Applications (including ERP, SCM, HCM, and EPM) and Oracle AI Data Platform - to modernize analytics, automation, and user experience. The SAP sidecar provides a pragmatic path to modern enterprise AI and automation without destabilizing SAP core processes, by combining Oracle Integration’s SAP connectivity with Fusion Applications integration and AI Data Platform’s governed data and AI services.

This SAP sidecar approach provides a clean, non-invasive extension layer that sits alongside SAP.

  • Connect to SAP using supported integration patterns (SAP BAPI/RFC/IDoc/OData).
  • Unify SAP and Fusion data into governed data products on AI Data Platform.
  • Power AI- and workflow-driven experiences that can write outcomes back to SAP and Fusion Applications.

The sidecar enables:

  • Cross-system process automation
  • Governed data products for analytics and AI
  • Agentic experiences grounded in trusted enterprise data
  • Safe write-backs into SAP and Fusion Applications
  • Strong auditability and operational control

AI Data Platform is designed to unify governed enterprise data with foundational models so teams can build an enterprise lakehouse, AI agents, and applications on an integrated platform.

You can address many business scenarios with this solution. For example:

  • Blocked Invoice Resolution: SAP invoice status is retrieved, Fusion supplier data is checked, and AI-driven remediation is proposed via AI Data Platform.
  • Supplier Onboarding: Harmonize SAP and Fusion Applications vendor data, enforce compliance checks, and create supplier records in target systems.
  • Order Promising: Pull demand and supply signals from SAP and Fusion Applications, identify constraints, and recommend reallocation or expedite actions.

Architecture

This solution describes at a high level the architecture of the SAP sidecar on OCI. The sidecar orchestrator coordinates SAP and Fusion Applications connectivity, curates trusted data products on AI Data Platform, and invokes AI automation to drive actions back into enterprise systems.

The following diagram illustrates the architecture at a high level:


Description of sap-sidecar-arch.png follows
Description of the illustration sap-sidecar-arch.png

sap-sidecar-arch-oracle.zip

This is the end to end flow within this architecture:

  • Trigger: A business event (for example, a new purchase order, blocked invoice, or supplier update) or a user request (for example, “Why is this invoice blocked?”) triggers the sidecar.
  • Orchestrate: The sidecar orchestrator routes the request to specialized agents and tools (such as SAP read/write, Fusion Applications read/write, data product retrieval, and policy checks).
  • Connect and Collect:
    • From SAP: through the Oracle Integration SAP Adapter (connecting to BAPI/RFC/IDoc), and/or the SAP S/4HANA Cloud Adapter (connecting to OData business objects).
    • From Fusion Applications: through the Fusion Applications REST APIs and/or Oracle Integration ERP Cloud Adapter/business events patterns.
  • Land and Curate: Data lands in OCI Object Storage (raw/bronze), is standardized and mastered (silver), then published to Oracle Autonomous AI Database as “gold” data products governed for downstream use.
  • Reason and Act: AI services and rule/policy logic produce actions (such as recommendations, approvals, and updates), which are written back to SAP/Fusion Applications through the same connectivity layer.
  • Observe: Every step is logged, traced, and auditable (data lineage and operational telemetry) including by Oracle Cloud Observability and Management Platform services.

The architecture deploys four agent layers:

  • SAP data integration agent
  • Fusion Applications integration agent
  • Data product and "gold medallion" governance layer Oracle AI Data Platform
  • Additional agents

SAP Data Integration Agent

The SAP data integration agent is a specialized agent responsible for interacting with SAP in a supported and configurable way. It has the following primary integration modes:

  • SAP Adapter (Oracle Integration) for SAP ECC / on-premises: supports business objects (BAPIs), function modules (RFCs), and IDocs (queued execution).
  • SAP S/4HANA Cloud Adapter (Oracle Integration) for SAP S/4HANA Cloud: supports discovery/mapping of business objects and invoke actions including query/create/update/delete via SAP S/4HANA OData APIs (basic authentication or client certificate-based).

This agent encapsulates SAP-specific details (credentials, endpoint conventions, payload formats, retries/error mapping) so all higher-level orchestration stays “clean core” and portable.

Fusion Applications Integration Agent

The Fusion Applications integration agent is a specialized agent responsible for integrating with Oracle Fusion Cloud Applications. It has the following primary integration modes:

  • Fusion Applications REST APIs to view/manage data in Fusion Applications.
  • Oracle Integration – Oracle ERP Cloud Adapter to integrate SaaS/on-premises apps with Oracle ERP Cloud, including event-based patterns and simplified service interfaces.
  • Bulk/analytics extraction (optional) using OCI GoldenGate Data Transforms with a BI Cloud Connector (BICC) connector to extract data directly from Fusion ERP Cloud for downstream analytics.

This keeps Fusion Applications integration consistent with SAP integration, by using agentic tools with clear contracts, governed authorization, and predictable error handling.

Data Product and “Gold Medallion” Governance Layer (AI Data Platform)

AI Data Platform is the sidecar’s data unification and AI enablement foundation. It provides an integrated environment that:

  • Unifies structured, unstructured, batch, and real-time data into an open connected platform.
  • Supports building AI agents and apps faster with integrated tools and agent frameworks.
  • Publishes trusted “gold” data products for analytics and AI workloads.

The governance layer performs these key roles in the sidecar:

  • OCI Object Storage: raw landings, extracts, documents, and API payload archives (bronze).
  • Autonomous AI Database: the “gold medallion” layer for trusted, query-ready data products used by agents and analytics.
  • AI Data Platform Workbench and Analytics: a governed workspace to build pipelines, semantic assets, dashboards, and reusable data sets for business teams.

Additional Agents (As Needed)

Because the sidecar is modular, you can add agents without changing SAP or Fusion Applications cores. For example:

  • Master Data Harmonization Agent: resolves vendor/customer/material identity across SAP and Fusion Applications; manages survivorship rules.
  • Exception and Root-Cause Agent: explains blocked invoices, failed integrations, and pricing mismatches; generates operator playbooks.
  • Policy/Controls Agent: enforces approvals, segregation-of-duties checks, and audit-friendly decision traces.
  • Notification/Case Agent: creates cases/tickets, alerts stakeholders, and posts outcomes into collaboration tools.

About Required Products and Services

This solution requires the following products and services:

  • SAP
  • Oracle Integration
  • Oracle GoldenGate
  • Oracle AI Data Platform Workbench
  • Oracle Fusion Cloud Applications
  • Oracle Autonomous AI Database
  • OCI Object Storage
  • Oracle Application Performance Monitoring

See Oracle Products, Solutions, and Services to get what you need.