Start with normal McKinsey interview preparation. McKinsey says no additional preparation is required for its AI-enabled interview pilots, and there is no publicly released avatar question bank you need to memorize.

If you still want relevant practice, concentrate on listening and speaking rather than learning a special script. Explain a recommendation in under a minute, invite a realistic client objection, and give a short evidence-based reply. Adjust your position only if the new information actually changes the analysis.

Try a practice exchange where a client says that a proposed expansion failed before. First acknowledge the concern; then ask what drove the earlier failure; then say which evidence you would investigate. Keep it conversational. This is a practice prompt, not a reported McKinsey question.

Ask a practice partner or coach for feedback on clarity and listening. Continue your usual case preparation as well: the McKinsey Solve assessment guide explains a separate digital assessment, but Solve practice is not a direct substitute for avatar client-conversation practice.

For a deeper explanation, see our McKinsey AI avatar preparation guide for practical exercises.

To focus your practice, read the discussion about the types of client questions and interactions to expect.

I like the idea of practising objections rather than memorizing another format.

I tried recording a one-minute recommendation and realised I use way too much jargon.