McKinsey to test candidates with AI‑simulated client conversations

McKinsey & Company will have some entry‑level hires spend ten minutes talking to an AI‑driven client simulation as part of its North American recruiting, according to the Wall Street Journal. The test, designed to mirror 80% of real consulting scenarios, will be administered to hundreds of undergrads this fall.

Key points

• McKinsey will test business‑analyst candidates with a 10‑minute AI client conversation.
• The simulation represents 80% of real consulting scenarios, per the firm’s design.
• Human interviewers will evaluate responses; the AI will not score candidates.
• Hundreds of undergrads in North America may encounter the tool this fall.
• The test is limited to select offices and is still in the testing phase.

What happened

McKinsey announced that, in a pilot rollout this fall, some entry‑level candidates will engage in a brief dialogue with a virtual client powered by artificial intelligence. The Wall Street Journal reports that the AI “aims to reflect 80% of the scenarios McKinsey consultants might encounter in their day‑to‑day work,” and that it is not intended to “mimic an unhinged client.” Human interviewers will sit in the room to assess how candidates handle the simulated situations. The pilot is slated for undergraduates seeking business‑analyst roles in a handful of North American offices, with the WSJ estimating that “hundreds” may participate.

The backstory

Consulting firms have long used role‑play and case simulations to vet candidates, and the rise of generative AI has provided a new tool to scale that practice. McKinsey’s move follows a broader trend, as other companies experiment with AI avatars in hiring. Hayden Field, a reporter who tested several such tools in a March video, noted that these systems are gaining traction across industries that value interpersonal assessment.

Why McKinsey is doing this

The firm seeks to standardize and accelerate its interview cadence, reducing the time human interviewers spend on repetitive scenario drills. By offloading the initial client interaction to an AI that can consistently present a wide range of cases, McKinsey can focus expert panelists on evaluating nuanced responses. The company also hopes to capture a richer data set on candidate behavior for future analytics.

Risks and trade‑offs

While the AI simulation offers scalability, it may introduce bias if training data is skewed, potentially disadvantaging certain applicant groups. Human oversight mitigates this risk, but the system’s reliance on scripted scenarios could miss emergent, real‑world complexities that a live client might reveal. Moreover, the pilot’s limited geographic scope means the results may not generalize across all McKinsey offices.

What builders and businesses should do

• Pilot AI‑driven client simulations in a controlled hiring cohort to measure impact on interview efficiency.
• Validate the AI’s scenario coverage against your firm’s typical client mix to avoid blind spots.
• Incorporate human evaluators to review AI‑generated interactions and flag bias or performance gaps.
• Monitor candidate experience metrics to ensure the simulation remains engaging and fair.
• Develop clear guidelines for when the AI should defer to a live client conversation.

The bottom line

McKinsey’s pilot will reveal whether AI can reliably replicate the breadth of consulting client interactions in recruiting. Success could prompt other professional services firms to adopt similar tools in the next hiring cycle.