A trust company that keeps its own record closed is asking you to do the exact thing it warns you against. So we don't.
Trust authorities earn recognition slowly. In the years before an insurer or a regulator formally points to your record, you need a different kind of proof — one a skeptic can check today.
Ours is simple: we run the governance we sell, out loud. Claire's own agents operate under GATE — and so does LoanCirrus, a regulated fintech we also own, held to the same governance as any other tenant (common ownership, disclosed). Every draft, approval, correction, and revert is recorded in the same tamper-evident ledger we'd put in front of your auditor. When we walk you through it, you are not watching a demo environment. You are inspecting a machine that acts — under governance you can check. And it acts in the domain GATE is built for first: regulated finance, where an action moves money and an examiner can demand the record. What you inspect is the real stakes we designed for, not a toy.
Claire's own sales and marketing agents operate entirely under GATE — draft-only, human-approved, self-improving. Our outreach, our pipeline moves, our content: all of it clears the gate before it ships, and all of it is recorded. If our own governance ever failed, the record would show it. That exposure is the proof.
LoanCirrus — a lending-technology company operating across 25+ countries under common ownership — runs its AI agent workforce under GATE, in a domain where audit, oversight, and accountability are not features but obligations. It is held to the same standard as any tenant, with no special treatment, and it is where GATE's evidentiary weight is tested against real regulatory stakes.
LoanCirrus and Claire share common ownership through their founder, Michael Claire. LoanCirrus is a governed GATE tenant — not part of Claire — and receives no preferential treatment of any kind. We disclose the relationship prominently and without being asked, because a trust company that presents a related party as arm's-length has already failed the only test that matters.
Governance you can't examine is just branding. These are the records GATE produces on our own installs — the same ones your deployment would generate.
The hash-chained record of who approved what, when, and why — the evidence that survives an adversary.
What humans edited and rejected, and how the agents improved because of it. Proof of getting better, not just getting used.
Privacy-preserving comparison across governed installs — how a roster measures against the field.
An agent's cleared record, issued as a verifiable credential a counterparty can check without taking our word for it.
A closed console can hide a bad week. A glass one can't. We accept that trade deliberately: the willingness to be inspected is itself the signal, and it's one a competitor selling a private black box cannot copy without taking the same risk. We'd rather earn trust the expensive way than claim it the cheap one.
Today, inspection happens in a guided walkthrough of the live record — not yet a fully public feed. Publishing a bounded, always-on view of how Claire governs its own agents is on our roadmap, and it's the direction glass points. We'll tell you what's live and what's coming, plainly, rather than imply more than is true.
Bring your hardest question about what an AI agent did and who stood behind it. We'll answer it from the ledger, on a live install.
Book a conversation How the record works