Zum Inhalt springen
All services
Connect and automate

AI integration

AI pays off with free text, scans and requests that cannot be captured in rules. We use it on the two-zone principle: the AI drafts, fixed rules decide what actually happens.

AI makes senseYes, but only where fixed rules are not enough, and always with a clear limit on what the AI may decide.

Sounds familiar?

Invoices and receipts arrive as PDF or photo and need to be read.
Enquiries arrive as free text and should be pre-sorted or answered.
Knowledge sits in documents and nobody finds it when they need it.

What you get

  • Document recognition with review
  • Assistants with your knowledge base
  • Agents with a fixed rule set
  • Labelling of AI content under Article 50 of the AI Act
  • Processing in the EU, no training on your data

How it works

01CheckDoes your case really need AI?a few days
02PilotWith real examples and a measurable goal2 to 4 weeks
03ExpansionSharpen rules, build into processesongoing

Does your case really need AI?

Three questions, one honest answer. Often the result is: fixed rules are enough.

Do inputs arrive as free text, scans or photos?

Are there cases that cannot be captured in fixed rules?

Can a person review the result before it counts?

Answer all three questions.

Technology

Azure OpenAI in the EU, Anthropic Claude, knowledge bases, rule sets

Billing

Pilot at a fixed price, then by scope and usage.

Frequently asked

No. We use providers and contracts under which your data is not used to train models.