The reported 2026 pilot is a short, interactive conversation with an AI-generated client, not another full case interview. There are public descriptions of the format, but I would be cautious about assuming every candidate receives identical prompts.

The avatar may ask follow-up questions or introduce information, while human assessors observe how you respond. McKinsey has described it as a realistic client interaction rather than an exercise built around trick questions.

For example, a client might raise an operational concern after hearing a recommendation. A useful response would acknowledge the concern, clarify what is driving it, and explain which information would change your conclusion. That is an illustrative practice scenario, not a leaked prompt.

If you get an invitation, follow the instructions from your recruiter. For preparation, practise explaining a recommendation to another person and responding to unexpected but reasonable questions. We should not treat anonymous accounts of the pilot as verified firsthand experiences.

For a deeper explanation, see the McKinsey AI avatar interview guide for the publicly reported format.

For a closer look at the conversation itself, read the discussion on what happens during the AI avatar interview.

I have only seen the reporting so far, so I would also like to hear how the interaction actually feels in person.

The fact that someone is watching while you talk to an avatar sounds stranger than the question itself.

I wonder whether the follow-up discussion with the assessor is part of the same ten minutes.