Early access · invite-only
User interviews,
asked the same way every time.
Send a private link. People answer in their own time by text or voice note. Your core questions are shown word for word; an AI interviewer adds short, neutral follow-ups within the limits you set.
No signup. We reply by email.
Verbatim by design
Code decides what comes next. The AI never rewrites your questions.
Async, text or voice
Participants answer when it suits them, on any phone.
Quotes you can trust
Only quotes that match the transcript word for word survive.
How it works
From guide to synthesis in four steps.
- 1
Write your guide
Core questions in English and Dutch, follow-up limits per question, allowed probe types and a time cap.
- 2
Share private links
Single-use, expiring, revocable. No public signup and no participant accounts.
- 3
People answer async
Text chat or voice notes. They can skip, stop, or delete their answers with a personal link.
- 4
Read the synthesis
Per-interview summaries, themes against your hypotheses, and verified quotes.
Built for rigour, not just speed
UX methods built in
Card sorts, tree tests, first-click, five-second and A/B preference tests, plus SEQ, SUS and UMUX-Lite scored for you. Easy warm-up questions settle people in first.
Verbatim core questions
The AI may only add a bounded probe, and every probe passes a leading-question guard.
Voice notes, transcribed
Hold a conversation the way people already message: short voice notes, transcribed automatically.
EU storage, GDPR by design
EU data storage, consent screen, retention limits and a self-service delete link. Names and contact details are removed before any AI model sees the text.
Verified quotes
Quotes the model can't match exactly to the transcript are dropped. Nothing invented.
Synthetic dry runs
Test your guide against AI personas (terse, rambling, Dutch-only, prompt injection) before inviting real people. Never counted.
Usability tasks
Coming soon: participants try tasks in your product while the interviewer asks how it went.
What you get back
A synthesis card for each study. Illustration with made-up example data, not real results.
Rainy-day outings
“I asked in the WhatsApp group and then just hoped it was open.”
- job: rainy-indoor
- anxiety: opening hours
- tool: WhatsApp group
Compared with per-interview SaaS
| Tool | Cost for a 30-min interview | EU data storage | Self-serve |
|---|---|---|---|
| Juno | $45 per completed interview | Processing may be in the US | |
| User Intuition | $30 (bring your own participants) | — | |
| Conveo | ≈$150 ($5/min, $5,000 prepaid minimum) | ||
| Outset · Listen Labs · Strella | Quote only, annual contracts | — | |
| Verbatim | Cents of AI compute (our estimate). Pricing to be announced. | Invite-only for now |
Published list prices, checked October 2026. They can change. Our own pricing isn't set yet.
Questions
Does the AI make up questions?
No. Core questions come from your guide, word for word. The AI only decides whether to ask a short follow-up of a type you allowed, up to your limit. Leading or yes/no probes are replaced by your standard probe text.
Where is data stored?
In the EU. Voice notes are deleted after 30 days by default. Personal details are removed before text is sent to an AI model, and the consent screen names the AI providers used.
Can anyone sign up?
Not yet. It's invite-only while the first studies run. Request early access and we'll be in touch.
Which languages?
English and Dutch today.
Does it replace talking to people?
No. It's for scale and consistency. Run a few live interviews alongside it to calibrate.
Run your next study async.
Early access is open to a handful of founders, PMs and small research teams.
Request early accessRequest access
Verbatim is invite-only while we onboard teams. Tell us a bit about what you'd test. Already invited? Sign in.