Lilli is McKinsey's internal AI platform — built on top of large language models and trained on the firm's proprietary knowledge base, frameworks, and client deliverables. Inside McKinsey, consultants use Lilli to accelerate research, synthesize information, draft slides, and structure recommendations.

In the interview context, Lilli acts as a simulated consultant resource. You're given a case prompt, access to the tool, and a limited window to work through a structured analysis. The interviewer isn't evaluating whether Lilli gives you the right answer. They're evaluating how you engage with the output — whether you can direct it, critique it, and build on it with your own structured thinking.

This is a fundamentally different skill from the classic "pen, paper, and silence" case interview. You're no longer just a solo problem-solver. You're a junior consultant working alongside an AI co-pilot.

---

Before you think about tactics, get the framing right.

McKinsey is not testing whether you can use a chatbot. They already know Lilli works. What they want to know is:

1. Can you direct AI with precision? Vague prompts produce vague outputs. Strong candidates arrive with a mental framework, then use Lilli to fill in targeted gaps — not to generate a starting point from scratch.
2. Can you evaluate AI output critically? Lilli will sometimes surface plausible-sounding but structurally weak answers. McKinsey wants consultants who catch that — not ones who nod along.
3. Can you synthesize and communicate under pressure? The final output needs to be yours. The recommendation, the logic, the so-what — that still comes from you.

This is exactly the model that top-performing consultants use in real engagements. The interview is a compressed simulation of Day 1 on a project.

---

The single biggest mistake candidates make in AI-assisted interviews is opening with the tool.

Don't.

Before you touch Lilli, spend the first 60–90 seconds doing what you'd do in any case: clarify the question, identify the decision to be made, and sketch your issue tree mentally (or on paper). What are the two or three critical questions that, if answered, would crack this problem?

Only then should you start directing Lilli — toward specific, bounded tasks. Examples of strong AI prompts in a case context:

• "What are the typical cost drivers for a B2B SaaS company operating at $200M ARR?"
• "Summarize the key factors that affect hospital patient throughput in the US market."
• "What frameworks are commonly used to assess market entry viability in Southeast Asia?"

These are surgical. They fill a gap in your analysis. They are not open-ended requests for Lilli to "solve the case."

---

Lilli is powerful, but it is not infallible — and McKinsey knows this. The interviewer may deliberately let the tool surface a flawed answer to see if you catch it.

After every Lilli output, run a fast three-point check:

• Is it relevant? Does this actually answer the sub-question I asked, or did it drift?
• Is it complete? Are there obvious gaps or missing considerations relative to the case context?
• Is it defensible? Could I walk an interviewer through this logic step by step without it falling apart?

If the answer to any of these is "no," say so out loud. Narrating your critical evaluation is not a weakness — it is the signal McKinsey is looking for. Candidates who silently incorporate bad AI output are the ones who fail.

---

One underused technique: once you've built your own hypothesis, ask Lilli to challenge it.

Something like: "What are the strongest counterarguments to a market entry recommendation for this company?" or "What risks or assumptions am I not accounting for in this cost-reduction analysis?"

This mirrors how senior consultants use AI in practice — not as a first-draft machine, but as a red-team partner. It also demonstrates intellectual confidence: you're not afraid to stress-test your own thinking. That's a quality McKinsey actively recruits for.

---

The final two minutes of an AI-assisted case are no different from a traditional case close. You need to deliver:

• A clear, direct answer to the question posed
• Two to three supporting reasons, in priority order
• One material risk or caveat the client should monitor

Lilli cannot do this for you — at least not in a way that will impress. The synthesis is where your judgment shows up. Candidates who read from Lilli's summary output in the closing are immediately identifiable as people who couldn't actually internalize the analysis.

Write your own closing bullets. Speak them in your own words. The AI was a tool. The recommendation is yours.

---

The gap between knowing this framework and executing it under live interview pressure is significant. You need reps.

This is exactly what BoardroomIQ's Case Simulator Tool was built for. It's an AI-native case practice environment that mirrors the format of AI-assisted interviews — giving you a live case prompt, a structured AI co-pilot to work alongside, and real-time feedback on how well you're directing the analysis versus deferring to it.

Unlike static case books or YouTube walkthroughs, the Case Simulator puts you in the decision seat. You get immediate signal on whether your issue tree is sharp, whether your prompts are precise, and whether your synthesis holds up under scrutiny.

If you haven't run through a simulated AI case before your McKinsey interview, you're leaving significant performance on the table.

For foundational case structure — the kind of framework fluency that makes AI-assisted cases dramatically easier — start with the BoardroomIQ Case Interview Learning Hub to build out your core toolkit before layering in AI practice.

---

A quick calibration point: McKinsey interviewers in AI-assisted rounds are not evaluating your typing speed, your familiarity with the Lilli interface, or whether you found a clever prompt. They're evaluating the same things they've always evaluated:

• Structured, hypothesis-driven thinking
• Quantitative reasoning and comfort with ambiguity
• Communication clarity under pressure
• Business judgment

The AI component is a new delivery mechanism for the same assessment. Candidates who understand this stop wasting prep time on "AI tricks" and instead double down on the fundamentals — which is the right call.

---

McKinsey's move toward AI-assisted interviews isn't a gimmick. It's a direct reflection of how the firm operates today. Consultants use Lilli on real engagements. The interview is testing whether you can do what a first-year analyst actually needs to do: direct AI intelligently, evaluate it critically, and synthesize the output into a clear recommendation.

The playbook is straightforward:

1. Structure before you prompt
2. Audit every output
3. Use AI to pressure-test your thinking
4. Own the synthesis and the close
5. Build your reps in a realistic environment before the real thing

The candidates who crack this format aren't the ones who are most comfortable with AI. They're the ones who are most comfortable with structured problem-solving — and smart enough to use AI as a lever, not a crutch.

Get your reps in. The interview clock doesn't wait.