Particle.news
Download on the App Store

McKinsey Pilots Lilli Interviews as It Scales Human-and-Agent Workforce

Some U.S. finalists are being asked to use the firm’s AI tool during assessments, signalling that AI fluency is now a baseline expectation.

Overview

  • CaseBasix reports that select final-round candidates are completing an AI interview using McKinsey’s internal tool Lilli to prompt, evaluate outputs, and apply judgment, with the firm not commenting publicly on the pilot.
  • The AI step runs alongside existing case and personal experience interviews, and guidance suggests weak use of the tool can hurt a candidate’s chances even though deep technical expertise is not required.
  • CEO Bob Sternfels says McKinsey now operates with roughly 25,000 AI agents alongside more than 40,000 human employees and plans to enable each employee with at least one agent within about 18 months.
  • The firm cites measurable productivity gains from agents, including about 1.5 million hours saved over the past year and 2.5 million charts generated in six months, with non‑client‑facing roles reduced by roughly a quarter as output rose around 10%.
  • Sternfels signals a hiring tilt toward liberal‑arts graduates for creativity and a shift toward outcomes‑based client work that ties fees to impact rather than traditional fee‑for‑service advisory.