Pioneering
the future of
enterprise AI.

What we believe

AI is rebuilding how enterprises work. We do the research in-house and turn it into products companies can deploy.

Our work

decide.sh

Decision models for enterprise workflows.

decide-1.1 reads a ticket, a document or a claim and answers typed questions about it, with a probability for every option. It works out of the box. decide.sh models also fine-tune on a team’s own decisions: 18 minutes of training on banking support messages lifted routing accuracy from 61.4% to 89.2%.

#1 among hosted APIs on the JevBench API leaderboard, v1.8.1.
Hosted APIOverall score
decide-1.181.1
Super Decider78.5
Myles-G 12B78.2
kettei-fat77.7
Jev 1.14reference70.2

Overall score out of 100 among hosted APIs, as the board shows it on October 30, 2026. decide-1.1 is listed, not yet ranked. Raw v1.8.1 results.

How we work

Research that ships.

Our models go from the lab to a public API. decide-1.1 is live on decide.sh.

Proof comes standard.

Every model you train on decide.sh includes an evaluation report, scored on examples it never saw.

Built for the enterprise.

Your data trains only your models. Run them on-premises or let us host them.

Consult with us.

Bring us the hardest workflow in your company. We’ll give you a straight answer on whether AI can take it on.

Or write to sales@cadentic.ai