What it is
Know-Your-Agent (KYA) is a proposed identity and verification framework for AI agents that shop, book, and pay on behalf of people. It borrows its name and logic from Know-Your-Customer (KYC), the decades-old banking requirement that a financial institution verify who its human customers actually are. KYA asks the equivalent question of software: is the agent making this purchase request a registered, accountable entity tied to a real operator and an authorized user — or an anonymous bot impersonating one?
Technically, KYA implementations rely on cryptographic proof rather than the heuristics merchants have historically used to spot bots (device fingerprinting, User-Agent strings, behavioral scoring). An agent presents a signed credential — reporting varies on the exact mechanism, with some accounts describing Ed25519 signatures or signed JWTs — that a merchant or payment network can resolve back to a registered operator and, ultimately, a human or business that authorized the transaction. A request lacking this chain of proof is treated as untrusted traffic.
The concept has no single owner. Individual payment companies have each built their own version: Visa’s Trusted Agent Protocol (introduced in 2025), Mastercard’s Verifiable Intent (plus a separate “Agent Pay for Machines” product for machine-to-machine transactions, launched June 2026), and Ant International’s Agentic Mobile Protocol (an open-source wallet-transaction protocol from April 2026). The startup Skyfire is credited by at least one source with popularizing the “Know-Your-Agent” label itself, around 2024, before it became a general term of art across the industry.
Why it matters for AI governance and narratives
KYA sits at the intersection of two narrative threads the observatory tracks: agentic AI’s push from advisory to transactional autonomy, and the governance vacuum around who is accountable when an autonomous system acts wrongly. Framing this as an identity-verification problem — echoing regulated financial infrastructure like KYC — is itself a strategic move by the payment networks: it positions agentic commerce as a compliance and trust engineering challenge that industry can solve through interoperable private standards, rather than one requiring new statute or regulator-led oversight. That framing is contestable. Unresolved questions reported alongside the announcement — liability when an agent transacts improperly, revocation of a compromised agent’s credentials, data-sharing terms between competing networks, and dispute-resolution mechanics — are exactly the questions regulators or consumer-protection bodies might otherwise claim as their own territory. Watching whether KYA remains an industry self-governance exercise or gets folded into formal regulatory frameworks is a signal worth tracking for how the broader contest over AI agent accountability resolves.
Key facts and dates
On September 9–10, 2026, Ant International, Mastercard, and Visa announced they had begun collaborating on an interoperable KYA framework, so that an agent verified once with one payment provider would not need to repeat verification elsewhere. Each company will continue running its own verification and decisioning process, but the three said they would work toward shared principles across three areas: cross-network operator traceability (linking an agent to a validated operator, cardholder, or organization), shared certification requirements (assessing agents against common security and behavioral standards), and continuous transaction monitoring (combining identity and payment signals over time). Named executives quoted in coverage include Visa’s Rubail Birwadker (global head of growth products), Mastercard’s chief digital officer Pablo Fourez, and Ant International’s chief information officer Jiang-Ming Yang, all describing interoperability as necessary for agentic commerce to scale. The initiative is reported to be proceeding through BuildFin.ai, an industry platform convened by the Monetary Authority of Singapore, alongside a related effort called SAFR (Safeguards for Agentic Finance at Runtime). Market-sizing claims attributed to McKinsey project that AI agents could orchestrate $3–5 trillion of global consumer commerce by 2030 — the commercial stakes driving the push for standardization. As of the announcement, KYA is a framework under development, not a deployed standard; no implementation timetable, pilot transaction volumes, or list of participating merchants and financial institutions has been published, and the accountability and dispute-handling questions noted above remain open.
Note on sourcing: this explainer draws primarily on wire-service and trade-press coverage of the September 2026 announcement, cross-checked across multiple outlets, plus one third-party technical explainer (eco.com) whose account of KYA’s origin and mechanics could not be independently verified against a primary technical specification. Readers should treat the cryptographic implementation details as reported rather than confirmed against source documentation.