--- title: "The McKinsey AI Interview: What It Really Is and How to Prepare" description: "Updated June 2026 · By Florian Smeritschnig, former McKinsey Senior Consultant If you have a McKinsey final round coming up, you have probably seen the headlines. McKinsey is adding an AI interview,..." url: https://strategycase.com/mckinsey-ai-interview/ date: 2025-12-11 modified: 2026-06-04 author: "Florian Smeritschnig" image: https://strategycase.com/wp-content/uploads/2025/12/mckinsey-AI-interview-scaled.jpg categories: ["McKinsey", "AI in Consulting", "Case Interview"] type: post lang: en --- # The McKinsey AI Interview: What It Really Is and How to Prepare *Updated June 2026 · By [Florian Smeritschnig](https://strategycase.com/about/), former McKinsey Senior Consultant* If you have a McKinsey final round coming up, you have probably seen the headlines. McKinsey is adding an AI interview, and half the internet is telling you to drop everything and learn prompt engineering. That advice is mostly wrong, and following it can cost you the offer. The McKinsey AI Interview is a new, currently non-evaluative pilot. You work live with Lilli, McKinsey’s internal generative-AI platform, to solve a case-style task while an interviewer watches how you prompt, judge, and synthesize. It does not test your AI or prompt-engineering skill, and at this stage it does not affect your offer decision. Your case interviews and the Personal Experience Interview still decide whether McKinsey hires you. I spent five years at McKinsey as a Senior Consultant and have coached hundreds of candidates into MBB offers, more than 200 into McKinsey alone. My MBA coaching clients in the US were among the first to go through this format, so this guide is based on what they actually reported back, not speculation. My verdict: prepare like a strong McKinsey candidate first, then add a light layer of AI fluency so the tool feels familiar. Overhauling your preparation for a pilot would be a mistake. ## **Key Takeaways** - The McKinsey AI Interview is a live session where you collaborate with Lilli, McKinsey’s internal AI, on a case-style task. It launched as a US pilot in January 2026 and expanded to select European offices later that same month. - It is currently non-evaluative. McKinsey uses it for testing and calibration, not scoring, and tells candidates so directly. - You are assessed on judgment, structure, iteration, and communication, not on technical AI skill or prompt engineering. - It applies to in-person final rounds at select offices. Virtual interviews do not include it at this stage. - [The McKinsey Case interviews](https://strategycase.com/mckinsey-case-interview) and the [McKinsey PEI](https://strategycase.com/mckinsey-personal-experience-interview-the-only-post-you-need-to-read/) remain the primary drivers of your offer. Do not shift preparation away from them. - The most common failure is treating AI as an answer engine instead of a junior teammate whose output you still own. ## **What Is the McKinsey AI Interview?** The McKinsey AI Interview is an interview component in which a candidate collaborates live with Lilli, McKinsey’s internal generative-AI platform, to work through a case scenario or exhibit. An interviewer observes how the candidate prompts the tool, evaluates its output, and turns it into a structured, client-ready answer. It is not a replacement for the case interview or the fit interview. It is a third component layered onto McKinsey’s existing final round at select offices. McKinsey is testing it the same way it tested the [Solve assessment](https://strategycase.com/how-to-prepare-for-the-mckinsey-digital-assessment) years ago: as a pilot first, long before it counts toward any decision. The framing that matters most is this: AI is the medium, not the subject. McKinsey is not checking whether you can write clever prompts. It is checking whether you can think clearly while a powerful tool sits next to you. ## **Where the AI Interview Fits in McKinsey’s Final Round** The AI interview shows up as an extra component alongside the traditional McKinsey case interview and the PEI, not as a global rollout. Adoption is limited to select offices, and McKinsey explicitly tells candidates the tool is not used for formal evaluation yet. Here is what my first MBA clients actually experienced. They completed two traditional interviews first: both with a case and a PEI story. The third session was different. An interviewer facilitated their work with McKinsey’s internal AI tool on case scenarios and exhibits. They described it as exploratory, less structured, and less guided than the formats they had drilled for. McKinsey communicated clearly that this was a test format, not a scored assessment. | Final-round component | What it is | Status (June 2026) | | --- | --- | --- | | Case interview(s) | A live business problem you solve with the interviewer | Core, evaluative | | Personal Experience Interview (PEI) | A structured behavioral interview built on your own stories | Core, evaluative | | AI interview (Lilli session) | A facilitated session where you collaborate with Lilli on a case task | Pilot, non-evaluative | Two qualifiers decide whether you will see it at all: - **In-person versus virtual.** If you are interviewing in person at a participating office, there is a high likelihood you are part of the trial. If your interview is virtual, you are not part of it at this stage. - **Region and timing.** Trial runs started in select US offices in January 2026 and reached select European offices by late January. Broader rollout is expected across spring and summer 2026, in line with how McKinsey has phased in past recruiting changes, [as reported by Fortune](https://fortune.com/2026/01/14/how-to-get-hired-at-mckinsey-ai-tools-liberal-arts-creativity/). Candidates I coach most often see this in generalist Business Analyst and MBA Associate rounds. If you are an experienced hire interviewing virtually, you are unlikely to encounter it right now. ## **How the McKinsey AI Interview Works in Practice** You complete a realistic consulting task by interacting live with the AI tool, and the process mirrors actual engagement work. Strong candidates treat Lilli the way they would treat a sharp but junior team member: useful, fast, and occasionally wrong. In practice, you move through a loop: 1. **Prompt** the AI with clear, focused questions tied to the problem. 2. **Review** the output for relevance, accuracy, and gaps. 3. **Refine** or re-prompt when the response is incomplete or generic. 4. **Synthesize** the useful pieces into a structured, client-ready answer. The single most important point: the interviewer evaluates how you work with the tool, not the tool’s output itself. Accepting whatever Lilli produces at face value is the fastest way to look junior. You are expected to show judgment, take ownership of the final answer, and steer the tool rather than follow it. ![Infographic titled “The AI interview loop” for the new McKinsey AI Interview showing a four-step circular process for using AI in a McKinsey-style interview: prompt, review, refine, and synthesize. The center message says “You stay in control,” emphasizing that the candidate owns the judgment while AI acts as a support tool, not the decision-maker.](https://strategycase.com/wp-content/uploads/2025/12/McKinsey-AI-Interview-1024x725.png) ## **What McKinsey Is Actually Evaluating** The interview assesses how you think, not what you know about AI. Four capabilities carry the weight, and they are the same ones McKinsey has always valued in a case. | Capability | What McKinsey watches for | | --- | --- | | **Judgment** | Deciding what to trust, adapt, or discard in Lilli’s output | | **Structure** | Framing focused questions and organizing outputs logically | | **Iteration** | Refining prompts to close gaps, step by step | | **Communication** | Explaining your reasoning and next steps clearly | Strong candidates stay calm when the AI returns something imperfect, then take ownership of the final answer. That mirrors real consulting work, where AI supports the analysis but never carries the professional accountability. If you can do that under mild time pressure, you are already most of the way there. ## **Why McKinsey Added an AI Interview** McKinsey has been open about AI’s place in how it now works. Lilli, its internal platform, [aggregates more than 40 of the firm’s knowledge sources](https://www.mckinsey.com/about-us/new-at-mckinsey-blog/meet-lilli-our-generative-ai-tool) to speed up research, synthesis, and knowledge access across engagements. The usage numbers explain why recruiting is adapting. According to McKinsey, Lilli now handles [more than 500,000 prompts every month](https://www.mckinsey.com/capabilities/tech-and-ai/how-we-help-clients/rewiring-the-way-mckinsey-works-with-lilli), and more than 75% of the firm’s roughly 43,000 employees use it monthly. When a tool is that woven into daily client work, it makes sense to watch how a prospective consultant handles it. As a result, consultants are now judged not only on independent problem-solving but on how responsibly they work with advanced tools. The AI interview lets McKinsey observe whether a candidate can: - Use AI to support structured thinking, not bypass it - Apply judgment when outputs are incomplete or generic - Keep ownership of decisions instead of deferring to the tool - Explain their reasoning clearly inside an AI-supported workflow In other words, the interview is a preview of the actual job, which is exactly what the case interview has always been. McKinsey has also signaled a broader hiring shift toward curiosity and creativity over narrow technical pedigree, [adding the AI task to graduate interviews alongside its push into AI agents](https://www.hrgrapevine.com/us/content/article/2026-01-15-mckinsey-goes-all-in-on-ai-with-interview-testing-workforce-of-20000-agents). ## **A Reality Check: This Is a Non-Evaluative Pilot** Here is the part most competitor guides bury or skip, because panic gets more clicks. The AI interviews are currently not evaluative. McKinsey positions them explicitly as testing, calibration, and fine-tuning, not scoring. Only candidates in select offices participate, and they are handed a McKinsey-issued laptop with Lilli access for the session. McKinsey is doing what it always does with a recruiting innovation: piloting it quietly before deciding whether to make it count. The Solve assessment followed the same path, starting as a low-stakes experiment before becoming a formal gate. That has a direct, practical implication. Do not over-interpret early exposure to the AI interview. Its presence in your round today does not mean it is deciding your offer today. The candidates who lose sleep re-engineering their entire preparation around a non-scored pilot are spending energy in exactly the wrong place. The first thing to do is simple: confirm with your recruiter whether the AI interview is part of your final round. It is not universal, and a 30-second question removes the guesswork. ## **How to Prepare for the McKinsey AI Interview** Preparation should be light, focused, and proportional to the stakes, which are currently low. You do not need access to McKinsey’s internal tools. Practicing with any publicly available AI platform, ChatGPT, Claude, or Gemini, builds more than enough familiarity. A practical routine takes very little time: - Ask goal-driven, specific questions instead of broad ones. - Read every output critically rather than accepting it. - Refine your prompt when the answer is vague or off-target. - Summarize the useful parts into a clear, structured recommendation. - Practice explaining your thinking out loud as you go. Tailor the effort to your situation: - **Business Analyst and MBA candidates** interviewing in person at a US or European office are the most likely to face it. Spend an hour or two getting comfortable steering a public AI tool through a business problem. - **Experienced hires and virtual interviewees** can deprioritize this almost entirely for now and keep their focus on the case and the PEI. Most importantly, do not move preparation away from cases and the PEI. Those remain the primary determinants of your offer. The fastest way to fluency on the fundamentals is structured practice, which is what StrategyCase’s [McKinsey Interview Academy](https://strategycase.com/mckinsey-interviews/) is built around. A few hours with a public AI tool is a sensible add-on, not a substitute. ## **Common Mistakes Candidates Make** Most difficulty in the AI interview has nothing to do with technical skill. It comes from using AI the wrong way, and both mistakes signal weak consulting fundamentals rather than weak prompting. **Mistake 1: Treating AI as an answer engine instead of a support tool.** Candidates who expect a finished answer ask vague, unfocused prompts and get generic output in return. When the result is weak, they freeze instead of iterating or redirecting the tool. The fix is to drive the conversation: ask a sharp question, judge the answer, and re-prompt with purpose. **Mistake 2: Presenting AI output uncritically.** Candidates relay what Lilli produced without synthesizing it, prioritizing what matters, or tying it back to the client problem. Many cannot explain why they kept one suggestion and discarded another. That inability to defend your choices is what reads as junior. Both behaviors point at the same root cause. The interview is testing judgment, problem structuring, and decision-making. AI is the medium through which those core skills become visible, which is why the candidates who already structure well tend to handle it comfortably. ## **What This Means for Your Consulting Career** The McKinsey AI Interview reflects evolution in consulting, not reinvention. The core skills, structured thinking, clear communication, and sound judgment, stay central. What changes is the environment those skills operate in. Looking ahead, AI will keep supporting research, analysis, and synthesis. Rather than replacing problem-solving, it will speed up how consultants gather information, test ideas, and structure insight. That puts a premium on two abilities: critically evaluating AI output instead of accepting it, and explaining your reasoning as clearly as you reach it. Speed without a defensible “why” will not be enough. The anchoring principle does not move: accountability stays with the consultant, not the tool. AI can assist the process, but the judgment, the recommendation, and the outcome still belong to the human in the room. Learning to work productively with AI is becoming part of professional readiness, the way learning to build a slide or structure an analysis already is. McKinsey is not alone here. Bain is reportedly introducing its own AI case interview in summer 2026, which you can read about in our [Bain AI interview guide](https://strategycase.com/bain-ai-interview/). Expect more of the MBB and Tier-2 firms to follow the same pilot-first pattern over the next year. ## **Frequently Asked Questions** ### **Is the McKinsey AI interview evaluative? Does it count toward my offer?** Not at this stage. McKinsey runs the AI interview as a non-evaluative pilot used for testing and calibration, and tells candidates so directly. Your case interviews and the PEI are what decide your offer. That could change as McKinsey expands the format, so confirm the current status with your recruiter. ### **Do I need to know prompt engineering or AI tools to pass the McKinsey interview?** No. The bar is baseline competence with AI, not advanced prompt engineering. McKinsey is watching your judgment, structure, and communication while you use the tool. A few hours practicing with a public AI platform like ChatGPT, Claude, or Gemini is plenty. ### **What is Lilli, and will I use the real tool in the interview?** Lilli is McKinsey’s internal generative-AI platform, which handles more than 500,000 prompts a month across the firm. In participating offices, candidates are given a McKinsey-issued laptop with Lilli access for the session. You do not need access to Lilli beforehand, and you cannot get it. ### **Is the McKinsey AI interview only for in-person interviews?** As of June 2026, yes. If you interview in person at a participating office, you are likely part of the trial. If you interview virtually, the AI component is not included for you at this stage. ### **Should I change how I prepare for McKinsey because of the AI interview?** Only slightly. Keep cases and the PEI as your core focus, because they still decide the outcome. Add a light layer of AI fluency, a couple of hours of practice steering a public AI tool through a business problem, so the format feels familiar rather than intimidating. ### **Which other consulting firms are adding AI interviews?** Bain is reportedly rolling out an AI case interview in summer 2026. Given how quickly McKinsey moved from pilot to broader trial, expect other MBB and Tier-2 firms to test similar formats over the coming year. ## **Related Guides** - [How to Prepare for the McKinsey Solve Assessment](https://strategycase.com/how-to-prepare-for-the-mckinsey-digital-assessment/) - [McKinsey Solve: Sustainable Future Lab Guide](https://strategycase.com/mckinsey-solve-sustainable-future-lab/) - [The McKinsey Technical Expertise Interview, Explained](https://strategycase.com/mckinsey-technical-expertise-interview/) - [McKinsey Keep in Touch Program: A Full Guide](https://strategycase.com/mckinsey-keep-in-touch-program-a-comprehensive-guide/) ## **The Bottom Line** Treat the McKinsey AI Interview as a forward-looking signal, not a reason to overhaul your preparation. It is a non-evaluative pilot today, built to watch how you think alongside a tool McKinsey’s consultants already use every day. The skills it rewards are the ones the case and the PEI have always rewarded: structure, judgment, and clear communication. So prepare like a strong McKinsey candidate first. Then add a thin layer of AI fluency on top, enough that Lilli feels like a familiar junior teammate rather than a surprise. If you can think clearly, apply structure, communicate your reasoning, and treat AI as support rather than a crutch, you are already prepared. Want a former McKinsey Senior Consultant to pressure-test your case and PEI before the real thing? [Book a 1-on-1 coaching session with Florian](https://strategycase.com/florian-coaching/) and walk into your final round ready for every part of it. --- ***About the author:** Florian Smeritschnig is a former McKinsey Senior Consultant who spent 5 years at the firm, conducted more than 2,200 candidate interviews through StrategyCase and other platforms, and has coached his candidates to 700+ offers at McKinsey, BCG, Bain, and other top firms. He is the founder of StrategyCase.com and the author of three consulting interview and career books: “The 1%: Conquer Your Consulting Case Interview,” “The 1%: Case Interview Workbook,” and “Consulting Career Secrets.”* Share the content![](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F)[](https://twitter.com/intent/tweet?text=The%20McKinsey%20AI%20Interview%3A%20What%20It%20Really%20Is%20and%20How%20to%20Prepare&url=https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F)[](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F)[](https://api.whatsapp.com/send?text=The%20McKinsey%20AI%20Interview%3A%20What%20It%20Really%20Is%20and%20How%20to%20Prepare%20https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F)[](https://strategycase.com/mckinsey-ai-interview/)[](https://reddit.com/submit?url=https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F&title=The%20McKinsey%20AI%20Interview%3A%20What%20It%20Really%20Is%20and%20How%20to%20Prepare)[](https://mail.google.com/mail/?ui=2&view=cm&fs=1&tf=1&su=The%20McKinsey%20AI%20Interview%3A%20What%20It%20Really%20Is%20and%20How%20to%20Prepare&body=Link:https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F)[](https://telegram.me/share/url?url=https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F&text=The%20McKinsey%20AI%20Interview%3A%20What%20It%20Really%20Is%20and%20How%20to%20Prepare)[](https://www.facebook.com/dialog/send?app_id=1904103319867886&display=popup&link=https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F&redirect_uri=https%3A%2F%2Fstrategycase.com%2Fmckinsey-ai-interview%2F)