Author: Taylor Warfield, Former Bain Manager and interviewer

Last Updated: March 24, 2026

 

McKinsey AI interview is a new final round component where candidates collaborate with Lilli, McKinsey's proprietary AI platform, to solve business problems in real time. It is currently being piloted in select U.S. offices for Business Analyst roles and is expected to roll out more broadly in 2026.

 

In this article, you will learn exactly what the McKinsey AI interview is, how it fits into the final round process, what skills are being evaluated, and how to prepare without overcomplicating your strategy.

 

But first, a quick heads up:

 

McKinsey, BCG, Bain, and other top firms accept less than 1% of applicants every year. If you want to triple your chances of landing interviews and 8x your chances of passing them, watch my free 40-minute training.

 

What Is the McKinsey AI Interview?

 

The McKinsey AI interview is a live, time-bound exercise in which candidates use Lilli to work through a consulting-style business problem. You prompt the AI, review its output, refine your questions, and synthesize findings into a structured recommendation. The interviewer watches how you collaborate with the tool, not just what the tool produces.

 

McKinsey began piloting this format in late 2025 for select final round interviews in the United States, primarily for Business Analyst candidates. According to reporting from the Financial Times and Harvard Business Review, McKinsey CEO Bob Sternfels has said the firm now operates with roughly 40,000 human employees and 20,000 AI agents. That ratio signals just how central AI has become to day-to-day consulting work at the firm.

 

The pilot is currently non-evaluative, meaning it is being used for calibration and testing rather than formal scoring. However, multiple sources indicate that McKinsey plans to expand the AI interview more broadly in Spring and Summer 2026 alongside its accelerated recruiting timeline.

 

The key takeaway: this is not a test of technical AI knowledge or advanced prompt engineering. It is a test of whether you can use AI the way a consultant would. Think of it as collaborating with a smart but imperfect junior teammate.

 

How Does the McKinsey AI Interview Fit into the Final Round?

 

The McKinsey AI interview appears as an additional component alongside the traditional McKinsey case interview and the Personal Experience Interview (PEI). It does not replace either of these assessments. Instead, it adds a third signal that helps interviewers understand how you work with AI tools in realistic scenarios.

 

Here is how the updated final round structure compares to the traditional format.

 

Component	Traditional Final Round	Updated Final Round (Pilot)
Interview 1	Case Interview	Case Interview
Interview 2	Personal Experience Interview	Personal Experience Interview
Interview 3	Additional Case or PEI	AI Interview (Lilli)
Primary Hiring Driver	Case + PEI performance	Case + PEI performance (AI is additive)

 

The case interview and PEI remain the primary hiring drivers. According to candidate reports, most of McKinsey's scoring weight still sits on structured problem solving and demonstrated leadership. The AI interview functions as a final practical filter, not the main gatekeeper.

 

For a complete breakdown of the McKinsey final round interview, including case types, PEI dimensions, and how to prepare for each component, check out our detailed guide.

 

What Does the McKinsey AI Interview Look Like in Practice?

 

In the McKinsey AI interview, you are given a McKinsey-issued laptop with access to Lilli and presented with a business scenario similar to what consultants encounter on real engagements. Your job is to use Lilli to explore the problem, gather insights, and build a structured recommendation.

 

Based on candidate reports and industry sources, the interaction typically follows this flow:

 

• You receive a business question or client scenario from the interviewer

 

• You prompt Lilli with clear, focused questions to explore the problem

 

• You review the AI output for relevance, accuracy, and gaps

 

• You refine your prompts when the first response is incomplete or too generic

 

• You synthesize the information into a structured answer and present it to the interviewer

 

The interviewer is watching every step. They care about how you frame questions, what you do with imperfect outputs, and whether you can explain your reasoning. Simply accepting whatever Lilli produces is exactly the wrong approach.

 

According to McKinsey, 72% of colleagues actively use Lilli, generating over 500,000 prompts per month, with consultants reporting up to 30% time savings on research and synthesis tasks. The AI interview mirrors this real workflow. If you can work productively with AI the way a consultant would, you are most of the way there.

 

What Is Lilli and Why Does McKinsey Use It?

 

Lilli is McKinsey's proprietary generative AI platform, named after Lillian Dombrowski, the first professional woman hired by McKinsey in 1945. It launched firm-wide in July 2023 and has become central to how consultants work.

 

Here is what Lilli does and why it matters for your interview.

 

Capability	What It Does
Knowledge aggregation	Draws from 40+ curated internal sources and over 100,000 documents and transcripts
Research and synthesis	Answers questions with synthesized insights, source links, and expert recommendations
Workflow integration	Accelerates slide creation, research, and synthesis with up to 30% time savings
Thought partnership	Helps consultants pressure-test arguments, anticipate client questions, and connect insights across workstreams

 

Lilli operates in two modes. The first scans McKinsey's internal knowledge base (called KNOW), which contains decades of sanitized project work, industry primers, and analytical frameworks. The second mode searches external sources. In a Fast Company article, McKinsey described the 11-month development process as riding the "struggle bus" because so much of the firm's knowledge lives in PowerPoint files that AI initially struggled to parse.

 

Why does this matter for your interview? Because McKinsey wants incoming consultants who are comfortable, thoughtful, and effective when using Lilli from day one. The AI interview is a preview of the job itself.

 

What Skills Does McKinsey Evaluate in the AI Interview?

 

McKinsey evaluates four core skills in the AI interview. None of them are technical AI skills. All of them are consulting fundamentals applied in an AI-supported setting.

 

Skill	What Good Looks Like	What Bad Looks Like
Judgment	Critically evaluates AI output. Identifies what to keep, adjust, or discard.	Accepts AI output at face value without questioning assumptions or gaps.
Structured thinking	Frames clear, goal-driven questions. Organizes synthesis logically.	Asks vague, unfocused prompts. Delivers unstructured responses.
Iteration	Refines prompts when output is incomplete. Improves analysis step by step.	Over-prompts randomly or gives up after one attempt.
Communication	Explains reasoning clearly. Articulates what was used, adjusted, or discarded and why.	Cannot explain thought process. Presents AI output without context.

 

In my experience coaching hundreds of candidates, the strongest performers treat AI interviews exactly like they treat case interviews. They define the problem before touching the tool. They stay in control of the analysis. And they own the final answer.

 

The DRIVE Method for AI Collaboration

 

I recommend using a simple framework I call the DRIVE Method to structure your approach to the McKinsey AI interview. It ensures you demonstrate all four skills McKinsey is looking for.

 

• D = Define the problem. Before touching Lilli, clarify the objective, constraints, and what output you need.

 

• R = Request structured output. Ask Lilli a clear, specific question tied to your problem definition.

 

• I = Iterate on gaps. Review what came back. Refine your prompt to fill holes or sharpen the analysis.

 

• V = Validate with judgment. Decide what to keep, what to adjust, and what to discard. Challenge assumptions.

 

• E = Explain your reasoning. Present your synthesized answer and walk the interviewer through your thought process.

 

This framework keeps you in the driver's seat. It prevents the most common mistake candidates make, which is letting the AI lead the analysis instead of leading it yourself.

 

How Should You Prepare for the McKinsey AI Interview?

 

Preparation for the McKinsey AI interview should be light and focused. You do not need an advanced technical plan. You need your consulting fundamentals to be rock solid, plus a thin layer of AI fluency.

 

Step 1: Confirm the AI Interview Is Part of Your Process

 

Not every McKinsey final round includes the AI interview. It is currently limited to select offices and roles. Before doing any AI-specific preparation, confirm with your recruiter whether the AI interview will be part of your final round. If you are interviewing in person at a U.S. office, there is a higher likelihood you will encounter it. Virtual interviews are less likely to include the AI component at this stage.

 

Step 2: Master Your Case Interview and PEI First

 

This is the most important step. The case interview and PEI are still the primary hiring drivers. You will not reach the AI interview unless you first pass these components. Roughly 20 to 30% of candidates pass McKinsey final round interviews, according to industry estimates, and the vast majority of that filtering happens on case and PEI performance.

 

If you want to learn McKinsey case interviews quickly, my case interview course walks you through proven strategies in as little as 7 days.

 

Step 3: Practice AI Collaboration with Free Tools

 

You do not have access to Lilli, but you can practice the exact same skills using any publicly available AI tool like ChatGPT, Claude, or Google Gemini. The tool does not matter. What matters is how you use it.

 

Here are five practice exercises to build your AI collaboration skills.

 

Exercise 1: Problem structuring.

 

Give the AI a business scenario and prompt: "Break this problem into 3 to 4 logical components without going into detail." Review the output. Are the categories MECE? Would you adjust them?

 

Exercise 2: Hypothesis testing.

 

Prompt: "What are the most likely drivers behind declining margins in the U.S. retail banking industry?" Evaluate which hypotheses are strong and which need data to validate.

 

Exercise 3: Iterative refinement.

 

Start with a broad question, then narrow it. For example, go from "What are the risks of entering the electric vehicle market?" to "What are the top 3 supply chain risks for a traditional automaker launching its first EV by 2028?"

 

Exercise 4: Critical evaluation.

 

Prompt: "Argue the opposite recommendation." or "What are we underestimating?" Practice identifying weaknesses in AI output and deciding what you would discard before presenting to a client.

 

Exercise 5: Synthesis.

 

Prompt: "Summarize this into a one-slide executive takeaway." Then rewrite the output in your own words. Practice explaining the answer out loud as if you were presenting to a CEO.

 

A strong workflow for any AI interview question is: Frame the problem first, ask one structured prompt, refine once, challenge the output, then synthesize into your own answer. You do not need many prompts. Three to five well-crafted prompts are better than fifteen unfocused ones.

 

Step 4: Practice Explaining Your Reasoning Out Loud

 

The interviewer will ask you to explain your thinking. Why did you ask that question? What did you notice in the AI's response? What would you do differently with more time? Practice narrating your thought process while you work through AI exercises. This is a skill you can build in a few practice sessions.

 

Having coached hundreds of candidates through McKinsey interview questions, I can tell you that the ability to articulate your reasoning clearly is what separates borderline candidates from offer-getters. This is true for case interviews, PEI, and the AI interview alike.

 

What Are the Most Common Mistakes in the McKinsey AI Interview?

 

Candidates struggle in the McKinsey AI interview not because they lack AI skills, but because they misapply consulting fundamentals when working with AI. Here are the seven most common mistakes to avoid.

 

Mistake 1: Treating the AI as an answer engine.

 

The biggest red flag is accepting AI output without questioning it. Lilli is a support tool, not an oracle. Candidates who copy and paste AI responses without adding their own judgment signal that they cannot own the analysis.

 

Mistake 2: Asking vague, unfocused prompts.

 

Prompts like "Tell me about the market" produce generic answers. Strong candidates ask specific, goal-driven questions like "What are the top 3 barriers to entry in the U.S. specialty coffee market for a new entrant with no existing retail footprint?"

 

Mistake 3: Accepting the first response without iterating.

 

AI rarely gives a perfect answer on the first try. The interviewer expects you to refine your questions. If the first response is too broad, ask a follow-up that narrows the focus.

 

Mistake 4: Over-prompting instead of thinking.

 

Some candidates fire off ten prompts in rapid succession hoping to stumble onto something useful. This shows a lack of structure. Step back, think about what you actually need, and ask one well-crafted question.

 

Mistake 5: Failing to explain your reasoning.

 

The interviewer cannot read your mind. If you do not explain why you chose a particular prompt, what you liked or disliked about the output, and what your next steps would be, they cannot assess your judgment.

 

Mistake 6: Ignoring gaps or weaknesses in the AI output.

 

Strong candidates proactively identify what is missing. They say things like, "The AI covered revenue drivers but did not address cost structure, so I would want to explore that next." This demonstrates the analytical rigor McKinsey values.

 

Mistake 7: Over-preparing for AI at the expense of case and PEI.

 

This is perhaps the most costly mistake. Candidates who spend weeks on AI preparation but neglect their case interview and PEI skills will not make it to the AI round in the first place. Get your fundamentals right first. AI preparation should take hours, not weeks.

 

Will Other Consulting Firms Add AI Interviews?

 

McKinsey is the first major consulting firm to formalize AI collaboration in its interview process, but it is unlikely to be the last. BCG has developed an internal AI tool called Deckster, and Bain has a tool called Sage. Both firms are embedding AI into day-to-day consulting workflows, which makes it reasonable to expect that AI fluency will eventually become part of their hiring processes too.

 

For candidates, this means AI literacy is quickly becoming a baseline expectation across all MBB firms. The good news is that the skills are the same regardless of the tool. If you can collaborate with AI productively using the DRIVE Method, you will be prepared no matter which firm introduces an AI component.

 

This shift also has implications for how consulting work itself is evolving. McKinsey CEO Bob Sternfels told Harvard Business Review that the firm is "migrating away from pure advisory work" toward an outcomes-based model. The tasks that junior consultants performed 30 years ago are now handled by AI. That means the bar for what incoming consultants need to do is shifting higher and faster.

 

For a complete overview of how the McKinsey interview process works from application to offer, including all interview rounds and question types, check out our step-by-step guide.

 

Frequently Asked Questions

 

Is the McKinsey AI interview currently being scored?

 

As of early 2026, the McKinsey AI interview is positioned as a non-evaluative pilot. It is being used for calibration and testing, not formal scoring. However, McKinsey has indicated it may roll out the AI interview more broadly if the pilot proves successful, at which point it could become evaluative. Treat it seriously even if it is technically non-evaluative today.

 

Do I need prompt engineering skills to pass the McKinsey AI interview?

 

No. McKinsey is not testing technical prompt engineering. They are testing whether you can ask clear questions, evaluate AI output critically, and synthesize findings into a structured answer. These are consulting skills, not AI skills. If you can do a case interview well, you already have the foundation you need.

 

Which McKinsey offices use the AI interview?

 

The AI interview has been reported in select U.S. offices, primarily for Business Analyst roles. It has not been rolled out globally. Candidates interviewing in person at a U.S. office are more likely to encounter it than those interviewing virtually. Always confirm with your recruiter.

 

Can I practice with Lilli before my interview?

 

No. Lilli is a proprietary McKinsey platform and is not available to external candidates. However, you can practice the same skills using any publicly available AI tool. The platform is not what matters. Your approach, judgment, and communication are what the interviewer evaluates.

 

How long is the McKinsey AI interview?

 

Based on candidate reports, the McKinsey AI interview is a live, time-bound exercise similar in length to other final round interviews, typically 30 to 60 minutes. The exact duration may vary by office and format.

 

Should I prioritize AI prep over case interview prep?

 

Absolutely not. The case interview and PEI remain the primary hiring drivers at McKinsey. AI preparation should be a light add-on after you have mastered your core interview fundamentals. Spending a few hours practicing AI collaboration is sufficient for most candidates.

Everything You Need to Land a Consulting Offer

 

Need help passing your interviews?

• Case Interview Course: Become a top 10% case interview candidate in 7 days while saving yourself 100+ hours
  
  
• Fit Interview Course: Master 98% of consulting fit interview questions in a few hours
  
  
• Interview Coaching: Accelerate your prep with 1-on-1 coaching with Taylor Warfield, former Bain interviewer and best-selling author

  

Need help landing interviews?

• Resume Review & Editing: Craft the perfect resume with unlimited revisions and 24-hour turnaround

 

Need help with everything?

• Consulting Offer Program: Go from zero to offer-ready with a complete system

 

Not sure where to start?

• Free 40-Minute Training: Triple your chances of landing consulting interviews and 8x your chances of passing them