Most of our work is under agreement and does not appear here. What follows is what we are able to describe, kept to what we can stand behind.
An operating system for university administration, deployed across more than eleven institutions.
Admissions, counselling and student operations run against one description of how the organisation works. It is the sector that taught us where the seam between engine and ontology belongs.
Programme bands and the escalation rule are read from your own admissions policy. The system does not decide eligibility — it decides who should.
Representative interface — real structure, client data replaced.
A QR-code quality-control system built for a vehicle assembly line.
Units, stations, checks and a definition of “passed” — the same four things as an admissions pipeline, in an entirely different vocabulary.
Tolerances come from the specification for this variant, not from an average. A unit is held at the earliest station where the fault is knowable.
Representative interface — real structure, client data replaced.
Two systems built end to end: one that runs sales outreach, one that turns account signals into a next action.
Both were engineered as products rather than as internal tooling, and both are ours end to end.
Every signal is a fact from your own systems, with the record it came from. The system does not predict churn — it says what it noticed and who owns it.
Representative interface — real structure, client data replaced.
We do not publish client results we cannot evidence, and we do not present a range from one engagement as though it were a benchmark. Where we quote a figure on a call, it is either measured on your workflow or labelled clearly as an estimate we would need to confirm.
On the call we put real numbers against your own operation — including, where it applies, the places this will not help.
Thirty minutes. We will tell you what we would build first, roughly what it costs, and where this will not help.