Hironautrecruiting operations with AI
From location to reporting in one system, and every campaign should learn from the last. Recognised as a research project by the German research allowance certification body.
Every campaign starts from zero
Jobs in mid-sized companies, especially skilled trades, are filled through ads someone builds, runs and reviews by hand. What a campaign learned then lives in someone’s head or a spreadsheet, not in the next attempt.
A loop instead of one-offs
Hironaut is designed to connect four steps into a closed loop: where the right people live, which creatives speak to them, how budget is allocated, and what follows for the next campaign.
The loop
Four stations, one loop
The loop runs on its own. Tap a station to pause it. Below, the barrier that keeps applicant data away from creative generation.
Location analysis
Where do the people who fit the role live, and how far is the commute?
An allowlist in the code ensures applicant data never reaches creative generation.
Why this is research
Three questions are open and cannot be solved with existing methods: How can creatives be described so a method reliably predicts what works from little data? How does context retrieval stay fast enough for a natural phone call? And how does a knowledge model that only invalidates facts fit the right to erasure?
No choosing between people
Hironaut is meant to record details, make first contact and book appointments. Who gets hired is decided by the employer. That is a design decision, not a limitation.
certified research project
stations in the loop
funding period
Privacy here is not a setting but architecture: an allowlist in the code defines which data can reach which part of the system at all.
Next project
THE CONCIERGE