Selected work
Finance operations
We built a controlled Xero submitter. It refused 1,740 duplicates before commit in year one.
Read the evidence
We build AI inside your business, until it ships.
A senior AI engineer takes one workflow from the first conversation to production. What we build runs in your business and stays in your hands.
The person on your first call owns what ships. Your repo · your IP · your runbook.
See how we workIn production.
Owned by your team.
Everything your team needs to own and operate what ships.
Accepted into operation
The work crosses the line. Your team keeps the controls.
Engineering
How the work moves
Every room we enter already has AI: a small one, barely used, on a stool in the corner. We raise it. We never replace it.
-
Map
Find how the work moves.
-
Build
Build one workflow in your repo.
-
Ship
Ship with evals and a runbook.
-
Run
Tune, train and hand over.
Selected work
A ledger that refuses to double-post
- Duplicate journal entries refused in year one
- 1,740
- Year-end clean-up
- Under a day
A controlled Xero submitter refused 1,740 duplicates before commit in year one and cut year-end duplicate clean-up from three weeks to under a day. The operation served 18,000 subscriptions. Applying the published sensitivity model’s per-refusal assumptions to those measured refusals yields $435,000 to $870,000 of annual debt, credit and resolution value at risk, not an audited client saving.
Read the caseA team built to be handed over
- Engineers transferred in-house
- 14
- Still in post after one year
- 12 of 14
We built and employed 14 engineers in India, embedded them in one client team, then novated all 14 in-house; 12 of 14 were still in post a year later. Using an approximate direct-hiring baseline of 90 elapsed days and a 31-day median, the sensitivity model estimates roughly 590 working engineer-days brought forward and $885,000 to $1.48 million of accelerated capacity, not the client’s commercial results.
Read the caseAlso built for: a collaboration-services provider · a creator-economy platform · a mobile-app studio
Teams, found and employed
When what you need is a team of your own: the build tells us what the role actually needs, then we help you shape it, find the people, employ them compliantly wherever they live, and train them for the role once it is defined. We build teams, not placements, so the shape is ten or more hires over six months, and they transfer in-house when you’re ready. Your IP, always.
How talent worksAbout
“We’ve been the founders in the room. We know what it costs to get this wrong.”
Gopal Patel – Founder and engineer. Chief technology officer at Auriens, where he built its early-stage ventures arm. He has structured cross-border transactions, worked in commodity broking, and created the by-appointment technology concierge service at Selfridges.
Jon Percy – Co-founder and lawyer. Brings the legal view in early, while decisions are still open. Fourteen years as a consultant general counsel to scale-ups; he now leads the legal function at an energy software company and a credit platform. Founder of esya.law.
Tell us the workflow that hurts.
Or the team you can’t hire. We’ll come back with a build plan, not a proposal.