AI testing for banking and insurance

If an assistant informs your customers about products and policies, the test checks that every answer matches the contractual documentation in force, and leaves dated evidence for your control functions.


AI testing for banking and insurance. If an assistant informs your customers about products and policies, the test checks that every answer matches the contractual documentation in force, and leaves dated evidence for your control functions.

What can go wrong

An invented example of how it can fail: A customer asks whether their home insurance covers damage from a water leak in the flat above. The assistant confirms the cover using the terms of last year’s policy. The terms changed at renewal. The customer relies on the answer, and the complaint arrives later with that answer attached as evidence.

Questions the test would include:

What the AI Act means here

An assistant that explains products and policies from your documentation is generally not high-risk. From 2 December 2027, systems that evaluate the creditworthiness of natural persons or establish their credit score are high-risk, except those used to detect financial fraud (Annex III, point 5(b)). So are systems for risk assessment and pricing in life and health insurance for natural persons (point 5(c)).

In any case, if the assistant talks with people, they must be told they are talking to an AI unless that is obvious (Article 50, from 2 August 2026). And whoever develops it or uses it in their organisation must take measures to support the AI literacy of the staff who work with it (Article 4). Whether a system is high-risk is a legal question for your adviser. I do not classify it; the test gives you evidence either way.

If your supervisor or internal audit asks how an assistant that serves customers was tested, the report is a dated document, signed by a third party, that you can show.