McKinsey has started providing “AI‑enabled laptops” in its entry‑level interviews.
I believe that single sentence captures the future of consulting—and, more broadly, the future of how we work.
Earlier this year, McKinsey Global Managing Partner Bob Sternfels appeared on the HBR IdeaCast and delivered a message that, frankly, every white‑collar professional—not just consultants—should reflect on twice. I’ve summarized it.
1. The interview setting has changed — it’s no longer about “prompts” but about “judgment.”
In final interviews for business‑school graduates at certain U.S. offices, candidates are given a laptop with McKinsey’s internal AI “Lilli” turned on. They work through a case study, but the evaluation focuses on how they query Lilli, how they critically interpret the output, and how they refine it into their own conclusions.
This is not a prompt‑engineering test. It’s an assessment of whether you can think like a consultant in the AI era.
In addition, McKinsey released a free AI case practice tool globally last April, tearing down the entry barrier of consulting coaching that used to charge $200–$500 per hour. Now anyone can practice case interviews without limit.
"Now, the ability to think with AI matters more than the value of human coaching" is the declaration.
2. Organizational composition has changed — of 60,000 employees, 20,000 are AI agents.
Sternfels' statement, verbatim:
"If you ask how many people McKinsey hires, the answer is 60,000: 40,000 humans and 20,000 agents. A year and a half ago, the same 40,000 humans were accompanied by only 3,000 agents. In the next 18 months, every employee will work with at least one agent."
In 18 months, the number of AI agents has increased more than sixfold, and soon we will have a 'one person, one agent' system.
I don't read this simply as "McKinsey is doing well with AI adoption." It's a signal that the pyramid structure of consulting itself is collapsing, because agents are filling the roles once held by the junior associate corps that used to create slides and tidy data.
3. The business model has changed — from 'hourly billing' to 'outcome accountability'.
I believe this is the most important change.
Sternfels says clearly in HBR:
"We are rapidly moving away from pure advisory work and hourly billing models. About one-third of our revenue now already comes from the 'outcome-underwriting' approach."
In plain terms, it's a model that ties our fees to how much the client actually earned or saved as a result of our consulting. We're becoming co‑owners rather than just advisors.
When you also consider the Financial Times' recent coverage of 'Project Acorn', the picture becomes clear. It is a compensation redesign that reduces the cash component of partner bonuses and increases the equity (equity) component by 3~5%p. As revenue volatility grows with AI and performance‑based pricing, it more tightly aligns partners' interests with the company's long‑term performance.
A firm that used to sell time is transforming into one that sells outcomes.
4. So what does this mean for us?
From McKinsey's shift, I read three clues.
① "We spent a lot of time" is no longer an excuse. Diligence is no longer proof of performance. When AI is at your side, a deliverable produced after 12 hours of work can be less valuable than one created in 2 hours with AI assistance.
② The ability being evaluated shifts from 'execution' to 'judgment'. Being good at prompting is just the basics. The ability to critically question AI's output, reinterpret it in context, and take responsibility to draw conclusions — this is the new core competency. This is exactly what McKinsey looks for in interviews.
③ Only those who can take on 'Outcome responsibility' will survive. The shift from hourly to outcome-based is not just a consulting story. In every role, the question "What outcomes are you responsible for?" will intensify. The ability to quickly produce outputs with AI while taking responsibility for the quality and impact of those results — this becomes the new standard for professionals.
Noover aims for outcome-oriented, performance-driven AX transformation.
If you have any questions or want to discuss AX transformation, feel free to message us anytime.