USA 
United Kingdom
McKinsey is now incorporating artificial intelligence into its recruitment for graduate talent and has built an AI "workforce" of 20,000 agents to support its human employees.
CaseBasix, a US firm that coaches candidates applying to strategic consulting companies, reported that the consultancy has added an AI element to select final-round assessments. Candidates are being asked to work with an internal tool called Lilli to complete applied exercises resembling real client scenarios.
It said the task involves prompting Lilli, reviewing the results, and exercising judgment to produce structured answers. The emphasis is on how applicants collaborate with the system to reach insights. The coaching firm described the approach as an AI interview and said the goal is to measure reasoning and communication rather than specialist AI knowledge.
Candidates are typically presented with a business question or scenario and rather than building their analysis from scratch, they use Lilli as a support function to search for information, shape thinking, and refine conclusions. Advanced prompting techniques are not expected, but applicants should demonstrazte that they can treat the technology as a productive thinking partner and articulate how they arrived at their recommendations, similar to the way consultants interact with junior colleagues.
Interviewees that have completed the process described the assessment as a test of judgment, analytical clarity, and collaboration with a tool rather than a technical exam on artificial intelligence concepts.
The AI segment sits alongside two traditional elements - problem solving and structured thinking, and personal impact, leadership, and values.
AI competence gains traction in selection
Artificial intelligence continues to rapidly expand into wider business functions. Microsoft said in 2024 that McKinsey would be an early adopter of its Copilot Studio initiative, which can deploy autonomous AI agents capable of handling tasks such as answering client requests and identifying sales leads. Participants in the same cohort included the law firm Clifford Chance and specialist retailer Pets at Home.
McKinsey chief executive Bob Sternfels recently told the Harvard Business Review that the company had a workforce of some 20,000 agents supporting its 40,000 employees, highlighting the scale of its commitment to using artificial intelligence.
He said: "When people ask me how many people McKinsey employs, my answer is 60,000: 40,000 humans and 20,000 agents. A year and a half ago we had the same number of humans but just 3,000 agents. In another 18 months I think every employee will be enabled by one or more agents. We’ll have a workforce that is human and agentic, and we’re going to have to navigate that.
"How does that change our model? We’re coming around to the conviction that we’re migrating away from pure advisory work, away from the fee-for-service model. We’re moving to more of an outcomes-based model, where we identify a joint business case with our clients, and we underwrite the outcome by tying our fees to the impact our work delivers for them. This aligns our interests with those of our clients a lot more. I think that’s the way of the future, particularly as we tackle ever more complicated questions about technology and enterprise change."
Wouter Durville, CEO at TestGorilla:, commented: “AI isn’t a “nice to have” in hiring anymore. McKinsey’s pilot is a glimpse of where things are heading: if AI is part of the job, it should be part of the assessment. We’re already seeing a clear divide between companies that experiment with AI and those that hire for it properly. The winners are building AI-first teams across whole functions, with people who are comfortable working with AI every day. In practice, that means job descriptions that call out AI tools and data fluency as core requirements, backed by skills-based assessment that tests judgment and the ability to challenge AI output – from junior hires right through to senior leaders.”
