PILOT
Pick one critical action. We prove your rules hold, or show you exactly where they break.
You pick one action your systems already perform, a wire transfer for example. We turn your rules into a check, block anything that breaks them before it runs, and seal every decision. At the end, your risk team re-checks everything themselves, without us.
HOW IT RUNS
- 01
Pick the action
One action type your agents or scripts already run — a transfer, an access grant, a deletion, a deployment — and the rules that govern it today.
- 02
Prove the policy
We encode the rules, search for sequences that pass every rule yet break the intent, and prove the corrected policy holds.
- 03
Enforce and seal
The gate runs in your environment, in front of the action. Every decision, allow or block, is sealed.
- 04
Verify without us
Your risk team or auditor re-checks the certificate and the records on their own machine.
What you bring
- The action type, and where it runs today.
- The rule as it is written now: limits, approvals, windows, holds.
- One person who owns that rule and can say what it was meant to say.
- A place in your environment where the gate can sit in front of the action.
What you keep
- The gate, running in front of that action.
- A sealed record of every decision, like the evidence pack.
- What the solver found in your rule, including where it leaks, like the lab shows.
- The readings we chose for your rule, and the limits of what was proven, in writing.
WHAT A PILOT WILL NOT TELL YOU
Three things, said before you ask.
A GOOD FIT IF
- An agent, API or script can already execute the action without a person clicking approve
- The rules exist on paper — limits, approvals, windows, holds
- Someone will be asked to show that those rules actually held
Tell us which action.
Leave a work email. We reply to set up a first call about the action and the rule behind it.
