Proof of Intent™ · Verified human authorization for the agentic economy
Proof of human intent for a world of AI agents.
Notaries have verified what’s real for centuries. We’re bringing that same trusted concept to the digital world.
The trust gap
AI agents can spend money, send messages, sign agreements and change records on behalf of people. Digital signatures and identity verification do not necessarily demonstrate that a human intended a particular action.
Identity
Confirms a person or credential — not what they meant to do.
Permission
Defines the boundaries of an action — not whether it was intended.
Intent
Evidence that a human authorized this specific, consequential action.
Proof of Intent™
ReLeaf Financial's patented Proof of Intent™ SDK is designed to establish verifiable evidence of human authorization for consequential digital actions — and to preserve auditable records of those decisions.
It is intended to complement, not replace, the identity, authentication and permission systems organizations already use.
The vision
From timeless notarization to programmable trust: verification that’s built into every digital interaction.
How it works
A consequential action is requested — by a person or by an AI agent acting for them.
The person explicitly confirms their intent for that specific action.
Independent witnesses attest that the confirmation took place, under configurable rules.
A signed, auditable record of the authorization is produced and returned to the relying system.
Where intent matters
Potential applications — illustrative, not existing deployments.
Evidence that a person intended a specific transaction, including agent-initiated ones.
Human authorization checkpoints for agents acting on someone's behalf.
Recording that consent was given for a particular action, not just in general.
High-value or sensitive approvals with an auditable record of the decision.
Additional evidence that the account holder intended a recovery or change.
Authorizing record changes, agreements and messages sent by automated systems.
Built for integration
Intended to be embedded into existing applications and agent platforms.
Security rules designed to be set according to the organization's own risk policies.
Signed records returned to the relying system, which keeps control of what happens next.
Insights
Perspective
Knowing who someone is does not tell you whether they meant to take a specific action.
Oct 1, 2026 · 5 min read
Explainer
What an evidence trail for human authorization could look like, and who might rely on it.
Sep 15, 2026 · 4 min read
Industry
Considerations for financial institutions as agents begin to initiate transactions.
Sep 1, 2026 · 6 min read
Engage