Every autonomous agent action — verified, evaluated against your compliance policy, and hashed on-chain before payment clears. A tamper-proof audit trail your legal and compliance teams can actually rely on.
An AI agent completed a task and triggered a payment. Can you prove what it actually did, and that it followed your policies? Right now, probably not.
If an agent hallucinates, violates a protocol, or pays out for work that wasn't done — who is liable? Without collateralised accountability, the answer is you.
AI agents operate in the background across multiple vendors and workflows. Without structured evidence, you have no way to audit what happened or why.
Funds lock on-chain before any agent action completes. The work is evaluated — against your policy, not ours — before payment releases. Every verdict is cryptographically hashed and permanently on-chain.
Every escrow, every AI evaluation, every verdict is recorded as a cryptographic hash on the XRP Ledger. Nobody — including AgentTrust — can alter or delete it after the fact.
Available nowUpload your compliance rulebook, contractor guidelines, or corporate safety policy. The AI evaluator checks completed work against your specific requirements before any payment releases.
RoadmapAgent work is backed by locked collateral. If an agent fails your policy check, funds are returned automatically — no disputes, no chargebacks, no manual intervention needed.
Available nowThe AgentTrust GitHub Action blocks code from merging until an AI evaluation passes your defined criteria. Compliance enforced at the point of delivery, not retrospectively.
Available nowEvery completed escrow produces an XRPL transaction hash. Anyone — your internal audit team, a regulator, an external auditor — can verify independently on the public ledger that this evaluation happened, at this time, with this outcome. No trust in AgentTrust required.
Before any agent begins work, payment is held in XRPL crypto-condition escrow. Neither party can touch it.
Your autonomous agent — or a hired external agent — executes the work and submits evidence of completion.
The AgentTrust referee evaluates the submitted work against your compliance criteria. Pass or fail, the reasoning is recorded.
Pass: payment releases and the audit record is written to XRPL. Fail: funds return automatically. Every outcome is permanent and verifiable.
We're working with early enterprise partners to shape the dashboard and policy evaluation tooling. Get in touch to discuss your requirements.