McKinsey & Company is incorporating artificial intelligence into its recruitment workflow by introducing an AI client simulation for entry-level candidates. As part of this new evaluation process, certain applicants are asked to communicate with an AI-generated avatar that assumes the role of a client during the interview stage.
This ten-minute exercise is intended to test how well candidates think quickly, address unexpected inquiries, and manage difficult dialogues. McKinsey is piloting the initiative this autumn with undergraduate students applying for business analyst positions across select North American offices, with hundreds of applicants expected to participate. Should the trial prove successful, the firm intends to expand the exercise to additional roles in the following year.
During the evaluation, candidates sit in McKinsey offices and speak to the avatar on a laptop while human partners and interview assessors observe the interaction from the room. The artificial intelligence system does not evaluate or score the applicants, nor does it make hiring decisions. Rather, human interviewers judge how effectively candidates listen, address client worries, and adapt when presented with new information.
The tool has been programmed to adopt various client personas, such as a professorial figure seeking detailed justifications for advisory recommendations. Designers of the pilot emphasize that the avatar is not engineered to behave unreasonably or pose trick questions, but rather to simulate roughly eighty percent of typical day-to-day consultant scenarios. Assessors evaluate whether applicants notice critical details and respond appropriately rather than blindly following a predetermined agenda.
This development arrives as various organizations test AI recruitment technologies, ranging from automated resume screening to anti-cheating software. While some employers remain hesitant due to regulatory shifts and potential bias concerns, McKinsey suggests that standardized avatars could enhance fairness by reducing the variability inherent in human interviewers challenging candidates.