What we hold ourselves to
Data stays where you say
On your servers, a private cloud or a national cloud.
Answers show their source
Staff can check where an answer came from.
A person decides
People approve decisions that affect rights or money.
Tested on real cases
Every model is measured on your own data before launch.
Logged and reviewable
Every question, answer and action is recorded.
How the rules work in a project
Before the build
We agree where data lives, who can see what, and which decisions stay with people.
During the build
The model is tested on your own cases. The results are written down.
After launch
Every question, answer and action is logged. You can review them.
When the model is unsure
The case goes to a person.
AI we have delivered
Private LLM
Language models running inside government networks
2deployments on client infrastructure
Document AI
AI that reviews documents for a mobile financial service
Mobile financial services provider
Machine learning
A dispatch engine that assigns orders with machine learning
foodpanda
AI route planning
A sales and distribution app with AI route planning and data bots
32%more sales volume in 6 months

Predictive AI
A platform that predicts customer service needs
62%fewer support calls
AI assistants
Web, mobile and social conversations in one AI-assisted console
55%faster response time


