McKinsey AI Client Interview: Entry-Level Candidates Face a 10-Minute Simulated Conversation

McKinsey & Company is adding an AI simulation of a client to its interview process for some entry-level candidates, according to The Wall Street Journal. Applicants will have a 10-minute conversation with the simulation, which the firm says aims to reflect 80% of the scenarios its consultants might encounter day to day. It is a small change in format, but it moves a core piece of consulting hiring from a human role-player to software.

What candidates will actually face

The reported format is simple. Some entry-level candidates will talk with an AI playing the part of a client for about 10 minutes. McKinsey describes the simulation as built to cover roughly 80% of the situations a consultant might meet in daily work.

Several details have not been disclosed. That includes which offices or candidate groups are involved, how the conversation is scored, and how much weight it carries in the final decision.

How this differs from the classic case interview

Consulting interviews have long centered on the case interview. An interviewer presents a business problem, and the candidate asks questions, structures the issue, and proposes a recommendation. In practice the interviewer also plays the client: they decide which information to hand over and how to react to a weak question.

Swapping that role for an AI changes who is on the other side of the table. A human interviewer adapts based on gut feel and experience, and different interviewers behave differently. A simulation can, in principle, be run the same way for every candidate. Whether it actually delivers that consistency depends on how it is configured, which has not been described publicly.

What the 80% figure does and does not tell us

The 80% number describes scenario coverage, meaning how much of a consultant's typical work the simulation is meant to reflect. It is not a measure of how accurately the AI judges candidates, and it is not a pass rate. McKinsey has not said how the coverage figure was calculated.

The 10-minute length is also worth noting. It is too short to walk through a full case from start to finish. That suggests, and this is an inference rather than a stated goal, that the exercise tests something narrower: whether a candidate can ask useful questions, listen to the answers, and adjust in a live exchange.

Why a firm might do this

There are a few plausible reasons, though McKinsey has only described the format itself.

• Consistency. A scripted-but-flexible AI can give each candidate a comparable scenario.
• Interviewer time. Human interviewers are senior consultants whose hours are billable, so moving part of the interaction to software could free some of that time.
• Signal about the job. A firm that expects its consultants to work with AI tools every day has an incentive to see how new hires handle conversational AI.

None of these is confirmed. They are reasons the move would make sense given how consulting recruiting works.

What it means for candidates

Preparation habits may need to shift. Memorized frameworks and rehearsed openings matter less if the counterpart responds unpredictably. Practicing live dialogue, including clarifying questions, summarizing what you heard, and revising a hypothesis mid-conversation, looks more useful than drilling polished monologues.

Candidates will also have to get comfortable talking to a machine as if it were a person with stakes in the outcome. That is an unusual skill to test, and it is unclear how much of it is about consulting ability versus comfort with the interface.

What it means for other employers

Many companies already use automated screening, such as one-way video interviews and online assessments. Those tools mostly filter applicants. A simulated counterpart that plays a work role is different, because it tries to observe on-the-job behavior rather than score answers to fixed prompts.

If McKinsey reports good results, other professional-services firms and large employers hiring for client-facing roles may try something similar. If candidates or regulators push back on fairness and transparency, adoption could slow. Concerns about bias in automated hiring tools are well documented, so how the scoring is audited will matter.

Takeaway

McKinsey's change is narrow, applying to some entry-level candidates and a 10-minute exchange. But it replaces the interviewer's role-playing with software at one of the most closely watched hiring processes in business. The open questions are scoring, consistency, and human review, and the answers will determine whether this stays an experiment or becomes a template.

Reference: https://www.theverge.com/news/1011654/mckinsey-ai-client-interview