McKinsey adds an AI client simulation to some entry-level interviews

The new exercise puts candidates in a conversation with an AI simulation of a client.[2] It is scheduled to last 10 minutes.[2] One possible scenario features an AI-generated mayor whose city is building a pricey new arena.[3] The applicant must offer a strategy in real time, and the mayor may introduce additional concerns or context as the candidate makes recommendations.[3] The setup tests how applicants respond as a client conversation develops.

McKinsey says the simulation is intended to assess how candidates think on their feet and handle conversations with clients.[1] The AI provides a consistent client persona, allowing assessors to focus on candidates’ listening, responses and ability to build trust.[3] The simulated client is not intended to imitate an extremely difficult or erratic client.[2] That distinction matters: the exercise is framed around responding to a client’s changing concerns, not managing an intentionally hostile exchange.

McKinsey says the AI does not score candidates; human assessors evaluate them.[2] Partners will observe candidate performance during the pilot, and the article says humans will make hiring decisions based on candidates’ experience with this type of problem-solving.[3] McKinsey told the Journal that the tool, which remains in its pilot phase, is meant to make hiring fairer.[3] That is the stated aim, not evidence in the reporting that the pilot has already produced fairer outcomes.

The tool is being tested in some North American offices this fall with undergraduates seeking business-analyst jobs; The Verge reported that hundreds of candidates may use it.[3][2] The business-analyst role has a reported starting salary of $128,000, and the article puts McKinsey’s acceptance rate between 0.5% and 1.5%.[3] McKinsey already uses a gamified assessment called Solve for most consulting roles, while its hiring guidance says tests or games may be required depending on the client-facing role.[4]