Updated 17 October 2026. This article has been revised to separate the policy timeline from the product argument. See the current AI Act timeline for the category-specific dates and primary reference.
A customer can ask a difficult question while the legal calendar remains unsettled. The useful response is a truthful account of the product, not a prediction about when enforcement will become inconvenient.
Answer the requirement you were actually asked
Clarify whether the customer is asking for a legal assurance, a contractual commitment or evidence of behaviour. These questions can sound similar in a procurement questionnaire while requiring different people and different answers.
For an illustrative request about intervention, demonstrate the relevant stop or approval mechanism. Explain what it can prevent and what it cannot reverse. If the customer needs a stronger capability, record the gap rather than broadening the claim.
Keep the conditions of the demonstration visible. A result from a test environment with limited tools may not establish the same behaviour in a wider deployment. Identify the configuration, the action and the conditions under which the test passed.
When the request falls outside current scope, decide whether to change the product, change the proposed deployment or decline the requirement. “Yes” is a very short word with an unusually long implementation tail.
| Decision | Evidence to request |
|---|---|
| Clarify | Assurance, commitment or evidence |
| Demonstrate | Current capability and conditions |
| Resolve the gap | Product or deployment decision |
What to do with this
A clear limitation can support a better commercial decision. A vague assurance postpones the disagreement until the customer has already depended on it.
Record the decision, the responsible owner and the evidence still missing. The decision-record example shows a lightweight way to do that. Where a claim concerns system behaviour, connect it to evaluation evidence. These are proposed product practices, not a finding that a system complies with the law.
Primary timeline reference: European Commission, AI Act. The operational recommendations above are my analysis.
