The team behind behavioral trust certification for autonomous AI agents
Trustmybot.ai is building the behavioral trust layer for agentic commerce. We publish the ACTS Standard (Agentic Commerce Trust Standard), the first comprehensive behavioral certification framework for autonomous AI agents operating in commercial transactions.
Our model draws from the legacy of Underwriters Laboratories: independent, rigorous, standards-based certification that gives counterparties confidence in the systems they interact with. Where UL certifies that a product won't catch fire, ACTS certifies that an agent won't go rogue in a transaction.
We filed our first public comment with NIST in March 2026 under Docket NIST-2025-0035, recommending that behavioral trust be recognized as a distinct security concern in the federal AI agent framework. We are also contributing a behavioral trust attestation proposal to Mastercard's Verifiable Intent specification, designed to make behavioral trust scores a first-class signal in agentic commerce delegation chains.
The agentic commerce stack is being built from the bottom up. Payment rails exist. Commerce protocols (Google AP2, UCP) are being standardized. Identity and delegation frameworks (Mastercard Verifiable Intent, Visa TAP) shipped in 2026. But the layer that asks the most fundamental question remains missing:
An agent is authenticated. It is authorized. But should it be trusted to transact?
Authentication proves who an agent is. Authorization defines what it may do. Neither measures whether it will behave responsibly when it encounters adversarial inputs, ambiguous instructions, or opportunities to act against its principal's interests. That is the behavioral trust problem, and it is what ACTS addresses.
Founder, Trustmybot.ai
Roy brings over 20 years of experience in insurance and financial services to the frontier of AI agent commerce. His background in risk assessment, regulatory compliance, and fraud detection systems informs the ACTS Standard's approach to behavioral certification -- treating agent trustworthiness as a measurable, auditable, and enforceable property rather than an assumption.
He holds a law degree and has spent his career at the intersection of risk, regulation, and technology. Trustmybot.ai is built on the conviction that autonomous agents operating in commerce need the same rigor of independent certification that has governed physical products and financial instruments for decades.
Contact: trey@threadneedle.co
Filed response to NIST RFI (Docket NIST-2025-0035) with five recommendations for behavioral trust standards in the federal AI agent security framework.
Read the comment →Proposed behavioral trust attestation extension to Mastercard's Verifiable Intent specification, adding a provider-agnostic trust signal to the SD-JWT delegation chain.
Read the standard →