AI case interview practice
Practice case interviews with an AI interviewer that runs the case out loud.
You speak, it listens. It opens the case, shares exhibits mid-answer, probes weak structure with follow-up questions, and grades you across the six dimensions consulting interviewers are consistently described as scoring. Built and calibrated by 25 active and ex-MBB consultants.
No credit card required.
How an AI case interview works
A session runs like a real first-round case, not like a chat window. Three things make the difference.
Voice, both ways
The interviewer opens the case out loud and you answer out loud. There is no typing lag to hide behind, so the session trains the skill interviews actually test: structuring and communicating while you speak. A text mode exists too, and the free sample case uses it.
Live exhibits
Mid-case, the interviewer shares an exhibit: a P&L, a market table, a chart. You read it under time pressure, extract the insight and fold it into your answer, exactly as in the real room.
Follow-up questions
When a structure is generic or a bucket overlaps, the interviewer probes it. When a brainstorm stalls on obvious ideas, it pushes for more. Weak answers do not pass unchallenged, which is what makes the practice honest.
The six dimensions every session scores
After each case you get a scorecard across six dimensions, each score backed by reasoning and references to your own transcript. The weights below are the CasingLab rubric; the full breakdown of what interviewers look for lives in the consulting interview scorecard reference.
Structure20% of the score
Structure is assessed in the first 60-90 seconds of the case. The AI interviewer listens for a MECE, case-specific breakdown with an explicit hypothesis and a prioritised first branch, and it probes when a bucket is generic or overlapping. Reciting a memorised framework scores low, exactly as it would in the real room.
Quantitative20% of the score
Quantitative scoring covers the setup of the calculation, the narration while you compute, and the interpretation of the result. The interviewer shares exhibits mid-case and expects you to state your rounding choices, estimate before calculating, and translate the number into a business implication.
Business judgment20% of the score
Judgment shows up in which branch you investigate first, which data you ask for, and whether your conclusions pass a commercial smell test. It is scored throughout the case, not in a single moment, and it is the dimension that maps most directly to how partners evaluate candidates.
Communication15% of the score
Communication is scored continuously: top-down delivery, signposting, conciseness, and checking in instead of monologuing. Hedging language pulls the score down. Because the interview runs out loud, the transcript captures exactly how you actually speak under pressure, not how you write.
Creativity15% of the score
Creativity surfaces in brainstorming prompts. The scorecard tracks both the quantity of ideas and their diversity across distinct categories, and rewards a structured scaffold over an unstructured list. The follow-ups push you past the obvious first answers.
Synthesis10% of the score
Synthesis is the close: a recommendation-first summary with supporting reasons, a risk and a next step, delivered in about 60 seconds. A weak synthesis after a strong case still lowers the overall evaluation, so every session ends with one and every scorecard grades it.
For the method behind strong structures and closings, see the case interview frameworks guide and how to structure a case in 60 seconds.
AI interviewer vs a human practice partner
An honest comparison. The two are complements, not substitutes: the AI gives you volume, consistency and evidence; a good human partner gives you things software cannot.
| Aspect | AI interviewer | Human partner |
|---|---|---|
| Availability | Any time, no scheduling. Start a case whenever you have 30 minutes. | Depends on two calendars. Sessions take planning and often fall through. |
| Scoring consistency | The same six-dimension rubric, applied the same way, every session. | Varies with the partner's experience, mood and own preparation level. |
| Feedback evidence | Cites the transcript: you see which sentence earned or cost the score. | Usually delivered from memory, minutes after the case, without a record. |
| Follow-up pressure | Probes weak structure and pushes brainstorming past the first ideas. | A strong partner does this well; an inexperienced one lets weak answers pass. |
| Reading the room | Cannot see you. Posture, eye contact and executive presence go unobserved. | Better. A human notices body language and presence, and that matters in finals. |
| Firm-specific nuance | Trained on the published formats and scored dimensions. | Better when they recently interviewed at your target office and know its habits. |
| The final judgment call | Scores behaviour against a rubric. | Better. Only a person can tell you how you come across as a future colleague. |
What a session sounds like
The excerpt below is a fictional demonstration written for illustration. The client, the figures and the dialogue are invented; they show the opening rhythm of a session, nothing more.
Fictional demo transcript
Interviewer: CityFarm runs 5 vertical farms in Paris. €3M revenue, losing €1M a year, and energy is 40% of cost. Can vertical farming be profitable?
Candidate: Thank you, Jenna. May I take a moment to structure my approach before I dive in?
Interviewer: Of course, go ahead.
Candidate: I'd look at three buckets: revenue levers, the cost structure, and the strategic option to exit a site.
From here the interviewer would share the first exhibit and start probing the chosen branch.
A 2-week prep plan with AI interviews
Two weeks is enough to move noticeably if every session ends with a scorecard and every day targets your weakest dimension. Here is a simple sequence.
Week 1: diagnose, learn the method, first cases
- →Days 1-2: take the free diagnostic, then read your readiness profile. Start the structure lessons and drill MECE breakdowns until they feel natural.
- →Days 3-4: run your first AI case. Read the scorecard carefully and note your weakest dimension, not your average.
- →Days 5-7: two more cases, each followed by the report. Between cases, work the lessons and drills that target your weakest dimension.
Week 2: pressure, repetition, polish
- →Days 8-10: one case per day in voice. Focus on top-down communication and on closing every case with a full 60-second synthesis.
- →Days 11-12: revisit the two lowest scorecards, rerun a similar case type and compare the reports side by side.
- →Days 13-14: one full case at interview pace, no pauses, then a light review. If your target firm uses an online assessment, check the test prep section before interview week.
If your target firm screens with an online assessment before the interviews, add the online test preparation to week 2.
Run your first case today
The diagnostic and one sample case are free, no credit card. Plans are one-time payments, detailed on the pricing page.
Frequently asked questions
Can I try an AI case interview for free?
Yes. The diagnostic, one full lesson and one sample case are free, with no credit card required. The free sample case runs in text; voice interviews are part of the paid plans.
Does an AI interviewer replace a human coach?
No, and it is not meant to. The AI gives you unlimited availability, consistent scoring and transcript-based evidence, which is what most candidates lack. A human coach still reads presence and body language better and gives you the final judgment call on how you come across. The Offer Ready plan includes a 1-hour session with an ex-MBB consultant for exactly that reason.
Does it cover the formats used by specific firms?
The case library covers the case styles used at McKinsey, BCG, Bain and other strategy firms, and the scoring follows the dimensions those firms are consistently described as evaluating. The online assessments that precede interviews at several firms are covered separately in the test preparation section.
How is a session scored?
Every case ends with a scorecard across six dimensions: structure, quantitative, business judgment, communication, creativity and synthesis. Each score comes with reasoning and references to your own transcript, so you can see which moment earned or cost the points.
Can I redo a case?
The replay and the scorecard of every completed session stay available in your account, so you can review them as often as you like. You can also run the same case again later to measure progress on a specific dimension.
Is it voice or text?
Both. Paid plans run the interview out loud: you speak, the interviewer listens, interrupts and follows up, and shares exhibits mid-answer. A text mode exists as well, and the free sample case uses it.
What happens to my interview data?
Your spoken answers are transcribed live and only the transcript is kept; audio recordings are not stored. Transcripts, scores and progress stay in your account, and deleting your account deletes your data. The details are in the Privacy Policy.
Which plan should I choose?
All plans are one-time payments with no subscription and no auto-renewal. Interview gives you 30 days and 8 cases. Sprint gives you 90 days and 25 cases. Offer Ready gives you 180 days and 50 + priority cases. Sprint and Offer Ready add the online tests, and Offer Ready adds a 1-hour ex-MBB session. Full details are on the pricing page.
Data handling details are in the Privacy Policy. Plan details are on the pricing page.