Artificial intelligence in the SAP environment — pragmatic and secure
How can AI be used sensibly in the SAP environment without migrating systems, handing over your data or building up new licensing models? We start with the existing system — delivering clear value for logistics, production and development.
Comprehensive AI consulting from a single source
CSBG delivers comprehensive AI consulting together with consiness GmbH & Co. KG — from the initial assessment through architecture and piloting to integration into your SAP landscape and ongoing operation. In this way we combine many years of SAP experience in logistics, production and development with our partner’s AI consulting spectrum.
We define the precise scope of services together with consiness, tailored to your starting point. More at www.consiness.de.
Secure, data-sovereign, without forced cloud
The technical foundation is designed to bring AI to your system — not your data to external systems.
AI without forced cloud
AI on the existing SAP system — on-premise, without migration, without mandatory BTP or Joule. Your data and control over it remain in-house.
RFC-MCP with SQL-Cache
Access to the SAP system via RFC, encapsulated in an MCP server. An upstream SQL-Cache makes analyses fast and relieves the production system — without restructuring the system and without permanent additional load on SAP.
Embedded AI in SAP — the ZAI concept
ZAI is not an off-the-shelf product for us, but an architecture concept that we tailor individually depending on the customer system: a dedicated SAP transaction with embedded LLM access, tool calling via curated ABAP building blocks, the authorization model adopted from the SAP standard, and the audit trail in the existing SAP logs.
Exactly this is what makes it policy-robust: where external model APIs are used, this happens in the user context and only as far as necessary — sensitive data stays in the system.
How ZAI is built
- Dedicated SAP transaction with embedded LLM access
- Tool calling via curated ABAP building blocks
- Authorization model adopted from the SAP standard
- Audit trail in the existing SAP logs
Why this is policy-robust
- No external agent plans API sequences from the outside
- The AI runs in the user context — tool calls are curated functions with the same authorization model as the clerk
- Read-only by default
- RFC filtering via a configuration table, clear whitelist
Suitable for
- Business users in day-to-day operations (FI, MM, SD, LE)
- 1st-level support, e.g. AuthPilot for authorization topics
- Scenarios in which the audit and authorization model must remain SAP-native
Where AI delivers concrete impact
Three fields in which AI delivers reliable value in the SAP environment today.
AI at runtime
AI directly in the running process: assistance for users, enrichment and plausibility checking of data, as well as answers based on your real SAP data — right where the work actually happens.
- Questions to the SAP system in natural language
- Plausibility and consistency checks within the process
- Assistance in logistics and production workflows
AI-supported development
AI as an accelerator in SAP development: from ABAP through interfaces (BAPI, IDOC, RFC) to testing and refactoring — embedded in CSBG’s many years of development experience.
- Code creation, review and refactoring (ABAP, interfaces)
- Faster onboarding into existing logic
- Support with testing and troubleshooting
AI for analysis, documentation & customizing
AI helps to understand what exists and to record it cleanly: system and process analysis, automatically generated documentation, and support with customizing decisions.
- Analysis of system, data and processes
- Generation and maintenance of documentation
- Decision support in customizing
AI potential in your SAP system?
Together with consiness, we take an honest look with you at concrete use cases. Without forced cloud, without a migration project.