Public data is everywhere, and most of it is messy and hard to use. We turn it into products, ship fast, and let buyers tell us what works.
See what's liveHow the studio works
We turn messy public data into products people pay for. AI does in days what used to take a team months.
One messy, valuable join that others can't easily assemble, with a buyer who already pays for something worse.
AI does the entity resolution, address matching, and extraction. Days, not months.
A working product into real hands, fast. The market is the only honest judge.
Keep what earns revenue, cut the rest, and reuse the assembly engine on the next one.
In the open
No vapourware. This board is the honest state of the studio. It moves as the work does.
Open for its own builds and forward-deployed client work. The opportunity map, eight open-data wedges scored across the UK and France, is the research the builds run on.
"Who really owns this building?" HM Land Registry gates title behind a per-title fee. We fuse the free pieces (corporate and overseas owners, land polygons, price-paid, Companies House and PSC) into one ownership graph. The join others won't do at scale.
Shipping log
Watch how we build, long before you decide to work with us.
The other half of the studio
When we're not building our own, we embed with industrial teams as forward-deployed data & AI engineers.