Built by a CFO who kept finding out too late.
Pursor was made by Andrew Playford, a fractional CFO working with founder-led consumer brands.
The job has a particular kind of bad week. An invoice arrives for an agency nobody mentioned. A new hire’s first payroll lands three weeks earlier than the forecast assumed. A supplier quietly moved to 50% upfront and the first anyone in finance hears of it is the payment run. None of it was hidden. Every one of those decisions was made openly, discussed in email, agreed by people doing their jobs properly. Finance was simply the last to know.
The usual fix is process — a form to fill in, an approval step, a standing item on a meeting agenda. It never holds. People are busy, the process is friction, and the commitment gets made in an email thread anyway. After enough years you stop believing the problem is discipline.
The commitments were always sitting there in writing. Reading all of it was the impossible part — and that is the part that recently stopped being impossible.
What Pursor refuses to be
A tool that reads company email could very easily become a surveillance product. That version would be easier to build and easier to sell, and it would be the wrong thing.
So the rule the whole product is built around is this: nothing reaches finance without the approval of the person it came from. If someone doesn’t answer, the report says a question is open with them — not what it was about. If Pursor picks up something personal, one reply deletes it. We are not the police, and an employee negotiating their own kitchen extension is nobody’s business but theirs.
Every part of the design follows from that, including the parts that make the product less impressive in a demo.
Early, and saying so
Pursor is new. Early customers get a person checking the work rather than a support queue, and a direct line to whoever can actually change the product. That is worth something now and it will not scale, which is rather the point of doing it while we are small.