We tried the smaller versions first.
- FirstA marketplace for people who were not looking. Getting someone hired through the person who already knew them worked well enough that vouching is still in the product, and still in the name.
- ThenAI screening, sold on top of whatever system a customer already ran. It helped, but it never changed how the day went, because it could see the candidate and nothing else.
- SoOnce you can say in seconds whether someone fits a role, the brief, the application, the interview guide and the shortlist all want rebuilding around that answer. That is a system of its own, so we built one.
Around a thousand recruiters have shaped it, four hundred of them in interviews over two months. What they said was the same every time: the tracking is fine, the help is thin, and what the team already knows never comes back out.
What we build on.
- Take the admin off the person, and leave the judgement.
- Drafting, screening, scheduling, summaries and translation are where the hours disappear. Clear them, and you get back the part of the job you were hired for.
- Every search should start where the last one finished.
- The expensive thing a recruiter produces is the thinking around a candidate: the read, the notes, the reason it was not right in March. Vouch keeps it with the person, so the next search starts from it.
- Understand the job before you judge the person.
- Two sales roles can share a list of requirements and be nothing alike. So the work starts with the role, talked through in detail, and everything after is written from that.
- An assistant that agrees with you is not helping.
- It asks where a brief is still vague, says what a rating has no evidence for, and marks what to test in the interview.