AI proposes. Your rules decide — before anything runs.
The proof includes its limit: what was shown, given which assumptions, and what is still trusted.
High Assurance AI is formally proved infrastructure for delegated execution, undertaken by Riley Betts as the Compliance Operator and as research.
The control problem starts when AI can change a record.
The problem starts when an AI can approve, submit, pay, or change a record. A draft is not the same thing. A wrong payment, or a wrong status, is a state you then have to unwind.
The Compliance Operator
The Compliance Operator is a policy written so it stays consistent, and a gate we have checked with mathematics against that policy. In the software that runs the AI model, it is the plugin that allows or blocks a tool call before the tool runs.
The check happens before the action. The evidence is what the gate recorded after it decided. Read how the Compliance Operator works.
Two ways to work with us.
The agentic governance and policy workshop, and designing and building the software. For one workflow, write what every acceptable action must satisfy, which assumptions that claim depends on, and what is still trusted. That is the authority map. It is what the Compliance Operator enforces.
Three papers, in the order they depend on each other.
Trust and Proof
The question is not whether a system is trustless. It is where the trust has gone: what was proved, relative to which assumptions, and what residue remains.
Read the paper →Goal-Relative Assurance
When an agent builds the path, assurance cannot mean “the approved process was followed.” The claim is that this trace meets this goal relative to this boundary, and that the boundary covers the floor the institution required.
Read the paper →Agentic Process Governance
An allow is not discharge. The institution, not the agent, pins the goal and the floor. Governance is the relation among the goal, the trace, the proof boundary, and the trust that remains.
Read the paper →All three papers, the glossary, and the patterns →
The Lean protocol stacks and Anchor are patterns we have researched and prototyped. The open source is the evidence. You reach them through the Lean-proved software offer.