Enterprise: ERP, CRM and SAP · Enterprise scenario · Proposed architecture
Creating a new customer or supplier (Master Data Governance)
From a new partner request to an active record: details are extracted from documents, duplicates are searched, legal, tax, bank and compliance checks run in parallel, a record is created or an existing one is extended, and the identifier is passed to every system.
Who it is for: Multi-company holdings and groups on a shared ERP
Systems: Request portal, CRM, SAP Business Partner data, Master data management system (if available), Procurement portal
Problem
Group companies create the same partner separately. Duplicate records pile up in the ERP, bank details are changed by email, and the CRM and procurement portal hold different identifiers.
How it works today
An employee emails company documents to the master data team. Details are entered by hand, similar records are searched by eye, and approvals stay in email threads. CRM and portal records are updated separately later.
What this scenario changes
The request starts in the portal. Details are extracted from documents, existing records are searched for exact and similar matches, and four checks run in parallel. For a potential duplicate a data steward decides, and a bank change is verified through an independent channel. The Business Partner record is created or extended with an idempotent write, and the identifier is passed to the CRM and the procurement portal.
The integration method is chosen based on the SAP version and the available interfaces.
Process stages
- Request and details: Get structured partner data from the request. Result: Name, tax ID, address and bank details
- Search in existing records: Find exact and similar matches. Result: List of candidate records
- Parallel checks: Legal, tax, bank and compliance checks at the same time. Result: Combined result of the four checks. Approval: Master data specialist (on a mismatch)
- Duplicate decision: Avoid creating a second record for the same partner. Result: A new record or an extension of the existing one. Approval: Data steward
- Bank change verification: Confirm a bank account change through an independent channel. Result: Confirmed bank details. Approval: Finance: person responsible for bank data
- Role approvals: Agreement from the finance and procurement roles. Result: Approved partner data. Approval: Finance and procurement
- Business Partner record: Create or extend the record exactly once. Result: Active Business Partner
- Cross-system sync: Pass the identifier to the CRM and the procurement portal. Result: The same identifier in every system
Participants
- Requesting employee
- Portal and CRM
- Integration layer
- ERP / SAP
- Data steward and master data
- Finance and procurement
- Partner company
The role of AI
Reads details from company documents and explains the differences between similar records for the data steward. It does not decide on duplicates, bank changes or approvals.
Rule-based automation
Exact and similar match search and the legal, tax, bank and compliance checks run on rules. The Business Partner write is idempotent through an operation ID, and on a sync error only the failed step is repeated.
Human approval and exceptions
For a potential duplicate the data steward decides, and when the bank account changes the person responsible for bank data in finance decides. Finance and procurement roles approve the record. When a check fails a master data specialist steps in, and when the retry limit is reached an integration operator does.
Frequently asked questions
Do you have a ready-made SAP integration?
This page shows the proposed architecture and a synthetic demo. The integration method is chosen in the project based on the SAP version, active modules and available interfaces. If there is a separate master data management system, the flow is adapted to it.
Does AI approve a bank account change?
No. The change is verified through the contact channel in the existing record, not the channel used in the request, and an authorized person in finance decides. AI only reads the documents and explains similar records.
What happens if the partner already exists?
No new record is created. The data steward reviews the similar record, and if it is the same partner, the existing record is extended with the new company code and role. The identifier does not change, and the same ID is passed to the other systems.