Rehearsa is AI case interview practice that talks back. A live interviewer runs a 20-minute consulting case with you on a shared whiteboard, interrupts you, pushes back, and scores you on the five dimensions interviewers use. Any hour, no partner needed.
Live session
Interviewer-led · 20 min · Phase 3 of 5, Analyze
Interviewer · Hold on. Costs grew 9%, but revenue grew 4% on 15% more rides. Which side is more alarming?
Every case is scored on the five dimensions interviewers use
Every session runs the way an MBB round runs. You structure, you earn the data, and you get graded on what you actually said.
It opens with the prompt, answers clarifying questions in a sentence, and makes you lay out a structure before it shows a single exhibit. Commit to the wrong branch and it pushes back, once, the way a partner would.
Sketch the tree on a shared whiteboard while you talk. The interviewer draws with you: it highlights the exhibit row that proves or kills your hypothesis and writes your numbers down as you say them.
Within seconds you get five scores with the written anchor for each, a phase-by-phase replay, your answer beside the model answer, and three next steps written about this interview.
Voice-first, push-to-talk. It interrupts you after 45 seconds without a headline, pushes back when you head the wrong way, and holds the data until you have earned it.
DriveCo grew rides 15% last year, but operating profit fell 22%. Where would you start?
I'd start on the cost side. Driver pay probably ballooned with the extra rides.
You jumped straight to costs. Revenue grew only 4% on 15% more rides. Which side is more alarming?
A dot-grid canvas you both draw on. You sketch the profit tree; the interviewer highlights, annotates and writes your numbers beside it.
No data until you have structured and chosen a branch. Ask for the cohort cut because you have a hypothesis, not because it is the next slide.
Five dimensions, scored 1 to 5 against written anchors, each with a note that references something you said in this case. Past 5% filler words, Communication is capped at 3.5 whatever the content was.
Five dimensions, tracked every case, so you know which rep to run next and what a 5 looks like on each one. The rules are the same every session.
45s
of rambling before it interrupts you
1
pushback per case, by design
3
hint levels per phase, every one counted
“Rehearsa is not a substitute for a human coach or a live partner. It is the reps between those sessions, the ones you would otherwise skip because nobody was free at 11pm.”
20 min
per live case
5
dimensions scored, 1 to 5
3
next steps per case
0
partners to chase
Start free. Pay when you want more cases than the free tier allows. Every plan runs the same interviewer.
The full interviewer, not a demo.
Start freeUnlimited reps, every one scored.
Go unlimitedFor consulting clubs and MBA programmes.
Talk to usFriends go easy and cannot score you against a rubric. Rehearsa runs the case, interrupts you, pushes back when you go the wrong way, and grades all five dimensions against written anchors. It is also available whenever you are.
No. The interviewer sits on top of a server-enforced state machine. It will not reveal data until you have structured. It pushes back once per case. It interrupts rambling. The scorer caps Communication when filler words pass 5%, and caps Communication and Structure again if the whiteboard went unused.
The free tier is a full case with the full interviewer and scorecard, not a demo. You pay only for volume.
Profitability and market entry today, interviewer-led, about 20 minutes each, tuned for McKinsey-style rounds. More case types are being added.
Yes. Switch to Type mode and the interviewer replies in text. Voice is the default because it is what the real interview tests.
Not yet. The shared whiteboard needs a laptop or a tablet in landscape.
Transcripts are kept for 90 days so you can review your sessions. Audio is not stored. Your sessions are yours; we never sell data.
Your next mock case interview is 20 minutes away, and it will not cancel on you.