Home / AI Bridge for SAP
// AI Bridge for SAP

Connect AI to SAP — safely.

AI Bridge for SAP is the connection layer between general AI clients — Claude Code, Codex, Gemini CLI — and your SAP ABAP and SAP BTP services. Access scope, landscape and authentication are controlled in layers, so AI can do real work grounded in your SAP data. Its core MCP server is published on GitHub under Apache-2.0.

// benefits

What AI Bridge for SAP changes

Investigation, verification and development work in SAP gets done inside the conversation with your AI client — no opening screens, recalling transaction codes, and copying results back and forth.

01

Less time spent investigating

Table reads, customizing checks, job and interface log analysis — all triggered in plain language. No screen navigation, no manual transcription of results.

02

No dependency on tribal knowledge

Nobody needs to remember transaction codes, table names or API specifications. First-level investigation that used to require a veteran can now be done by the person who owns the task.

03

One session across systems

ABAP systems, BTP services and non-SAP systems are reachable from a single AI session, so cross-system checks — matching an S/4 document against an interface log — finish in one pass.

04

Controlled access, with an audit trail

Three access scopes — full, reference-only and development-only — combine with per-landscape control over DEV / QAS / PRD. Every call is written to the audit log.

05

Minimal change on the SAP side

Remote-enabled function modules and BAPIs can be called without per-service web service configuration, so you can start without modifying the existing system.

06

On-premise, RISE or cloud

Connections run through a backend on SAP BTP, so on-premise, RISE with SAP and cloud services all behave the same way from the client.

// approach

How we roll it out

Never a big-bang rollout. We prove the value on a small scope first, then widen it together with the access-control design. The standard path is five steps.

STEP 01

Assessment

We map the target systems and connection paths (on-premise / RISE / cloud), the intended use cases, and who needs which level of access. This is where we draw the line between what AI does and what people decide.

STEP 02

Backend setup

We build the connection backend on SAP BTP, Cloud Foundry and configure destinations and authentication (XSUAA). All actual SAP communication and security enforcement happens there — the AI client never holds SAP credentials.

STEP 03

Security design

Each user gets an access scope (full / reference-only / development-only), and every destination is tagged DEV / QAS / PRD. Modifying operations and function module calls against production are denied by default. We also define key issuance, revocation and audit log operations.

STEP 04

Pilot

A small group runs real use cases so benefits and gaps surface early. Binary installation and distribution of connection settings are standardised at this stage.

STEP 05

Rollout and ongoing operation

We widen the scope department by department and system by system, and keep supporting audit log monitoring, additional tools and version updates.

// coverage

Systems we cover

From core SAP business systems to cloud services and the non-SAP systems around them — all reachable from a single AI client.

SAP S/4HANA

Function module / BAPI calls, table reads, Open SQL and DDIC access, OData V2 / V4, REST and SOAP services, add-on development and transport management.

SAP Ariba

Procurement and sourcing data, read and integrated through the published APIs.

SAP Concur

Expense and travel request data, read and integrated through the published APIs.

SAP Cloud Identity Services

Identity Authentication (IAS) admin and SCIM operations, plus Identity Provisioning (IPS) job execution and job logs.

SAP Integration Suite

Cloud Integration (CPI) monitoring and audit logs — enough to run first-level analysis on integration errors.

SAP Business ByDesign

Master and transactional data, read and integrated via its OData / SOAP services.

JIRA

Read, create and update issues over REST API v3, so findings from SAP land straight on a ticket.

SmartDB

Workflow and request data over REST API v3, matched against what actually happened in SAP.

JP1

Job scheduler status and logs, reconciled with SAP-side batch processing in one pass.

Beyond the above, the server covers BTP services including Cloud Foundry, Cloud Transport Management (cTMS), Build Work Zone, Datasphere, SAP Analytics Cloud, Cloud ALM, Alert Notification Service, Build Process Automation and Forms Service by Adobe. The current list of systems and available operations is published and kept up to date on GitHub.

Start with the systems and the use cases.

The core MCP server behind AI Bridge for SAP is public on GitHub. The backend it needs to actually run — setup, connection configuration and operation — is delivered as a consulting engagement. Happy to talk it through.