01
The Challenge
A traditional survey gives you a snapshot: one moment in time, captured after the problem has already happened. Feedback scatters across spreadsheets and email threads. And without real anonymity, people don't answer honestly, so the data quietly lies.
We wanted something better: a tool that tracked sentiment continuously, protected the people answering so thoroughly that not even we could unmask them, and found the signal in the noise on its own. Good enough to put our own name on.
02
How We Worked
We designed and shipped Signals end to end. A visual signal builder with five question types and reusable templates. Profile questions that segment results without breaking anonymity. On top, a real-time analytics dashboard with presentation mode, and AI insights that cluster open-text answers and track sentiment across time.
The privacy model is the centrepiece, because honest answers are the whole point. Responses are anonymised with one-way hashing and the email-to-hash mapping is never stored: a response can't be traced back to a person, even by us, while duplicate submissions are still blocked. An analytics threshold hides results until enough responses are in, so individuals can't be identified inside small teams. Anonymity isn't a policy promise; it's how the system is built.
03
The Outcome
Signals is live at signals.tech, launched with Software NI as partner. Pricing is outcome-based: free to 100 responses a month, then pennies per response, with no seats and no annual contract, so a team starts free and only pays as feedback scales.
It's also proof of how we work. The advice we give clients about production-grade software isn't theoretical. Signals runs on the same standards, and we design it, ship it, and operate it ourselves.