McKinsey Plans AI Client Interviews for Some Entry-Level Applicants

A simulated client, with people assessing

McKinsey plans to have some entry-level job candidates speak with an artificial intelligence (AI) client for about 10 minutes during interviews. Infobae and The Verge describe the planned exercise in coverage citing The Wall Street Journal.

The fall 2026 pilot covers undergraduate business-analyst applicants at some North American offices and could involve hundreds of candidates, according to the supplied reporting. Candidates will hold the conversations in McKinsey offices while partners or assessors observe.

The AI avatar—the simulated client candidates speak with—does not score or evaluate them, Infobae reports. People retain that responsibility, making the conversation an exercise for human observation.

McKinsey's AI client exercise puts listening to the test

Recruiting co-leader Blair Ciesil says heavily rehearsed candidate stories prompted the effort to assess applicants more directly. McKinsey partner Saleem Zafar says assessors look for candidates who listen to important information, act on it and avoid imposing their own agenda.

For applicants in the pilot, that means showing how they respond during a conversation, alongside any prepared account of their skills. The stated focus puts weight on using what the client says to guide a response.

Zafar says the simulation represents about 80% of consulting situations and does not center on trick questions or an out-of-control client. That percentage is his characterization of the exercise.

From using an AI assistant to speaking with an AI client

Infobae reports that some McKinsey candidates have already used Lilli, the firm's internal AI assistant, to solve a case since January 2026. In that exercise, partners observe how applicants question the assistant.

The two exercises place AI in different roles. With Lilli, applicants use an assistant to work through a problem. In the client simulation, they respond to a client while human assessors observe their listening and judgment.

Expansion to other roles in 2027 depends on the pilot working. But Infobae reports that McKinsey has not disclosed how it will measure success or whether it will compare participants with candidates who did not use the avatar. Whether the exercise improves hiring assessments is therefore still an open question.