Describe your decision in minutes. Like a thoughtful partner, Rehearsals asks you questions, helps you narrow in on the audience you care about, and gets you thinking about what outcome you're measuring and what you want to see.
Recorded interview moments from real people represent the individuals behind each Rehearsals AI twin.
Predict marketing outcomes by simulating AI twins of real people
The only simulation platform with its own foundational model and first-party panel. Predict human decisions with 94% accuracy, measured against public benchmarks.
Built and
backed by
A new era of decision making with simulated worlds.
What you can predict
Why Rehearsals
- Make decisions, not surveys.
- Rehearsals agent will ask questions to understand your decision. We'll handle the rest, but you can change anything.
- Vertically integrated and verifiable.
- See the data behind a decision, end to end. We'll even follow up with an individual person to prove a point.
- On your time, starting today.
- A carefully designed experience that makes incredibly complex technology easy for people to use. Or let your agent do all the work.
An agent that does less talking and more doing. Ask complex questions. We'll spin up an entire background virtual machine and download gigabytes of simulation data just to answer your question as fast as possible.
Instantly verified from the real world. Rehearsals runs hundreds of validation tests comparing humans, AI twins, and real-world outcomes. Every time you run a Rehearsal, we draw on our vast library to give you a score and evidence that's verifiable and understandable.
Full transparency because you should check. Rehearsals is the most transparent simulation platform. You can see the conversations, stimuli, and even foundational base data to figure out how a decision was made.
A simulation is only as good as its outcome. Record the outcome to help fine-tune your account. We don't train on any customer data. When you're done, get a thoughtful list of what-ifs and variables you could change in your next Rehearsal.
94%Agreement between AI twins and their real humans.
AccuracyHUMAN-TO-TWIN VALIDATION
We prove accuracy on your questions.
With your permission, we measure alignment and accuracy between AI twins and humans by running human-to-twin validation studies that take no more than a day or two.
Explore validationYour question
Which product concept would you most likely buy?Predicted. Then observed.
Four results from the Replica Performance Whitepaper, shown only against what happened in the real world.Read the whitepaper (PDF)Most subscribers stayed put.
Twins forecast that 84% would absorb a $3 increase, versus 94% observed.
91% accuracy
OutcomeObservedTwins
Accept increase94%84%
Switch to ads<1%14%
Travelers kept booking.
Twins predicted 93% would still book after a 10% fare increase, versus 91% observed.
98% accuracy
OutcomeObservedTwins
Still book91%93%
Skip9%7%
Matched the iPhone mix.
Which iPhone twins chose, compared with the published U.S. sales mix.
97.6% accuracy
OutcomeObservedTwins
Pro Max27%29.8%
Pro25%27.1%
Picked the winning ad.
Twins chose Ad B, the creative that delivered 3.53x ROAS versus 2.52x for Ad A.
72% picked the winner
OutcomeROASTwins
Ad A2.52x28%
Ad B3.53x72%
Human Behavior Benchmark.
The first public and transparent human behavior benchmark.
Coming
| Model | Score |
|---|---|
 Rehearsals twins | 94.0% |
 Claude Opus 5 | 78.7% |
 Gemini 3.8 Flash | 77.5% |
| Gemini 3.1 Pro Preview | 76.9% |
 GPT-5.6 Terra | 76.0% |
 Grok 4.20 (Non-Reasoning) | 71.6% |
AI twins of real individual people, not synthetic populations.
AI twin of a team member
Live panel markets
Dots are illustrative and do not represent population size.
- Video verified
- Participants are interviewed on camera.
- 100% human reviewed
- Our team verifies every interview.
- Continuously refreshed
- Participants average 5+ interviews every quarter.
Custom recruitment, included at no extra charge.
We’ll find your niche audience, interview them, and build their twins in about a week.
The Rehearsals team.
From Google DeepMind, Google Labs, Microsoft, Databricks, and Wayfair. Backed by Andreessen Horowitz.
Meet the team
Built to be trusted.
- Works where your team works.
- Run Rehearsals natively in Slack, Teams, Claude, and ChatGPT.
- One twin, one person.
- Every twin is built from one real person and answers as that person.
- Your data never trains our models.
- Customer data, briefs, questions, and reports are never used to train any AI model.
- Enterprise ready.
- SOC 2, GDPR, SSO, ongoing penetration testing, and third-party-verified deletion.
- Transparent by design.
- A clear commercial model, published methodology, benchmark results, and a human-to-twin validation on your own questions.
- Keep your rehearsals forever.
- You own your questions, cohorts, and reports with a perpetual license. Rehearsals retains the underlying twins, panel, and models.
QUESTIONS, ANSWERED
Everything people ask before the first Rehearsal.
How the twins work, how we prove accuracy, what the panel looks like, and how engagements are structured.
Read every answerA Rehearsal is a business decision simulated with your customers. You bring a question, Rehearsals runs it with a defined audience of AI twins, and your team gets a decision-ready report back in under an hour.
GET STARTED
Start with the decision you need to make.
Run your next creative, pricing, launch, or retention decision with AI twins of real people.