Skip to content

AI that understands SAP: ground rules in the platform itself

An AI assistant does not know the quirks of SAP. When the platform itself supplies the rules, pitfalls and working patterns, the AI does not have to guess.

The problem

AI assistants are happy to write ABAP and call SAP functions. But SAP is full of quirks that are hard to find documented: limits in the editor, functions that only work in a certain order, pitfalls that only show up in production. An AI that does not know this makes the same mistakes again and again.

Supplying the ground rules

Our solution: the platform tells the AI how it works. With every answer it supplies:

  • hard rules: what is never allowed;
  • known pitfalls: what went wrong before and why;
  • working patterns: how to do it right;
  • diagnosis: how to recognise and fix an error.

On top of that there is a manual for the AI and a collection of working methods built up from real incidents.

Why in the platform and not in a document

A document gets forgotten or goes out of date. Rules that the platform supplies with every answer are always there and always current. Every new lesson becomes a new rule, and every AI session benefits from it straight away.

Where it is heading

The next step is a platform that an AI assistant can "talk" to directly, so a developer can have an SAP integration built from their own editor, within the same rules.

What to watch out for

  • Record every lesson from an incident as a rule for people and AI alike.
  • Never let the AI place anything in SAP without a check; the system always verifies.
  • Keep the rules short and concrete: "don't do this, do this instead".

Would you like advice on AI and SAP? Request a free advice conversation.

Updated on

AI

Personal data in the AI cloud: who is watching?

A conversation with an AI service in the cloud is not private by default. A case in Canada shows what a provider can see and share. What does that mean under the GDPR, in the United Kingdom and in Switzerland?