Our mission
Make it possible to decide, with evidence, whether a specific AI agent is safe to deploy in a specific context, and to keep checking once it's live.
About us
We've spent careers making software prove itself. Now we do it for AI agents.
Every serious piece of software gets tested before it goes live. AI agents, the systems that read your email, update your records and move your money, mostly don't. Security teams are asked to approve them on a vendor's word. We think they deserve evidence.
Independent by design
We don't build or sell AI models, so we have no reason to go easy on one. Our only product is an honest answer.
Built in the UK
A UK company, building tools that UK organisations can run on their own terms, mapped to the guidance their auditors already use.
[Founder photo]
The founder
Abdalla Yusuf
Abdalla is a QA lead who has built test-automation frameworks from scratch for web and mobile platforms used by [a large user base], cutting regression time and testing costs along the way. [Add a sentence on why agent assurance, in your own words.]
The engine behind Olyptic (a calibrated AI judge, a test runner and CI gates) was first built to test AI assistants against UK law. It's now the foundation for testing agents that act.