McKinsey & Company is adding an AI-simulated client to its interview process for some entry-level candidates. According to The Verge, citing The Wall Street Journal, some undergraduates applying for business analyst roles in North America will have a 10-minute conversation with an AI version of a client this fall. Human assessors will be in the room to judge how they handle it.
Key takeaways
- The test is a 10-minute talk with a simulated client.
- It is being trialled for undergraduate business analyst candidates in some North American offices.
- Hundreds of candidates may take it, according to the WSJ.
- The AI does not score candidates; humans do.
- The simulation aims to reflect 80 percent of the situations consultants meet day to day.
What is McKinsey testing?
Consulting interviews have long used case studies, where a candidate works through a business problem with an interviewer. The new step adds a conversation with an AI client meant to reflect common day-to-day scenarios. The WSJ reported that it is not designed to "mimic an unhinged client". The point is to see how candidates listen, ask questions and explain ideas.
Why are employers doing this?
Employers want to see skills in action rather than rehearsed answers. Live simulations are harder to game with AI tools than written tests or take-home tasks. They also reflect the real job, where consultants work with clients and AI tools at the same time. Other companies are building AI avatar interviews too, and Greenhouse found that 48 percent of US job seekers took some form of AI interview in the past year.
The tool is still in testing, so McKinsey may change or drop it depending on how this fall goes. Candidates who get it should treat it as one stage among several, not a replacement for the rest of the process.
Who is affected?
For now, a limited group: undergraduates applying to McKinsey business analyst roles in certain offices. But large employers often copy each other's hiring methods, especially in consulting, banking and tech. Graduates applying to competitive schemes should expect more live simulations and fewer purely question-and-answer interviews.
How should candidates prepare?
- Practise structured thinking: clarify the problem, ask questions, then propose a plan.
- Listen more than you talk. Clients, real or simulated, reward people who understand their needs.
- Speak plainly and summarise often. Assessors are watching how clearly you communicate.
- Do practice conversations out loud with a friend or a chatbot, but prepare your own reasoning.
- Read our guides to common interview questions and virtual interviews.
What does this mean for interviews generally?
Hiring is splitting in two directions. Some employers are adding AI steps like this one. Others are bringing back in-person rounds to stop cheating, as our report on deepfake candidates explains. Either way, the skill being tested is the same: thinking clearly in real time.
Source: The Verge, 14 October 2026.
Keep this insight useful.
Share it with your network, save it for later, or create a signed affiliate link.
Creating or clicking a link does not create earnings. Only fraud-reviewed outcomes can allocate an existing eligible paid-ad-funded monthly pool.
