Research into personal AI agents that can work together safely.

Personal AI agents will increasingly deal with agents acting for other people. Valeway is independent, UK-based research into which safeguards still hold when nobody sees the whole exchange: a trust layer that decides in code, and escalates only what code can't decide. The goal is agents that people, organisations and public services can depend on when they meet.

What we're working on

Rules enforced in codeResearching permissions that are checked in code before an agent acts.
Safe agent-to-agent exchangeWays for agents acting for different people and organisations to work together without overreaching.
Open, testable researchMethods and findings published so others can examine and build on them.

Coming soon: a research write-up and open repository.

Get in touch

Research by Sebastian Vale. For enquiries, write to hello@valewaytechnologies.com.