Q: What is the Lilli test and how does it fit into McKinsey’s interview process? A: The Lilli test is a pilot AI interview where candidates are asked to use McKinsey’s internal tool Lilli to research, structure, and draft a response under time pressure. It may appear as an additional step alongside the case interview and a personal experience interview for candidates in the U.S. and North America. Q: What key skills does McKinsey evaluate in the AI interview? A: McKinsey will measure prompting and iteration, reasoning, communication, collaboration, and judgment. The McKinsey AI interview guide emphasizes testing whether candidates can prompt the AI, challenge its output, and convert notes into a clear, structured response. Q: How should I prepare before taking the Lilli test? A: Practice with any enterprise-style AI to build habits like scoping, citing sources, and versioning prompts, and get comfortable moving from broad prompts to focused ones. Use a steady prompting framework—role, goal, constraints, and format—and sharpen human skills such as MECE structuring, creativity, math sanity checks, and ethics. Q: What step-by-step playbook should I follow during the Lilli session? A: Restate the problem and confirm the target audience, set a brief that specifies what you need and the format, then draft an outline and drill deeper on the top two levers. Pressure-test the output for counterarguments and risks, check evidence or assumptions, prioritize 2–3 high-impact moves, and package a concise recommendation with risks and next steps in 60–120 seconds. Q: What are the main do’s and don’ts for the AI interview? A: Do use Lilli to explore, compare, and structure responses, cite limits and assumptions, and translate analysis into actions, owners, and timelines. Don’t copy raw AI text as your final answer, invent data or references, or rely on outside AI during assessments, since McKinsey prohibits that use. Q: How can liberal arts candidates show creativity and judgment in the Lilli test? A: Turn creativity into distinct, structured options across lenses such as customer, operations, product, and policy, and use short, relevant client narratives to make analysis stick. State what the model cannot know, call out ethical guardrails like privacy and fairness, and add human judgment to assumptions and risks. Q: How should I demonstrate outcome-focused thinking when using Lilli? A: Link each recommendation to measurable metrics like revenue lift, cost savings, churn reduction, or cycle-time cuts, and show a path to value with quick wins (30–60 days), mid-term builds (1–2 quarters), and scalable bets (6–12 months). Propose a test with one pilot, one KPI, one owner, and one deadline, and explain how outcomes-based work aligns incentives when relevant. Q: What practice drills and scoring methods does the guide recommend? A: Use the 15-minute drill: pick a short case, run an outline, dive deeper on two levers, perform a risk scan, and deliver a six-bullet summary while time-boxing analysis and packaging. Self-score on clarity, structure, evidence, judgment, and impact using a 0–2 rubric to track improvement.