The situation
Soluzione Latte's vets visit dairy farms all day. What they learned on those visits ended up spread across notes, lab PDFs, milking robot data, ultrasound files and herd exports, and most of it had to be typed in again by hand. There was no single place to see how a cow or a farm was really doing, and the farmer got their report at the next visit.
What we built
We made the visit itself the data source. The vet, the stablemaster and the farmer each wear a small microphone that records the morning, even without signal; then a model that knows veterinary language transcribes it, and a sentence like "cow 315, mastitis, this treatment" becomes a proper record for that cow. Lab results, milking robots, ultrasound files and herd exports all flow into the same database, built around the individual cow.
On top sits a dashboard for the vets only, with health, treatments, lab outcomes and what illness really costs in discarded milk and medicine. Vets get a brief before the visit, the farmer gets a clear PDF shortly after, quotes are filled in from the visit and the price list, and an AI co-pilot points out animals at risk, always as a suggestion the vet decides on. It was all tested on real pilot visits before going live.
The result
Today nothing is typed in twice: what is said in the stall ends up in the system, and the farmer has their report the same morning.