/ 01Platform
Identity primitives, not a bolt-on gateway.
AID, Wallet Auth, Transaction Guard, Bot-vs-Agent detection, and merchant verification — composable primitives your existing stack plugs into. No rip-and-replace.
See the platformWhat we build
Three layers. One trust stack.
We organize around what merchants and developers actually have to ship: verification primitives, the risk problems they solve, and the revenue outcomes they unlock.
/ 02Pre-Dispute Network
Catch agent disputes before they become chargebacks.
A side-channel for agent transactions: signed-trace evidence travels with every alert, and a 24-hour response SLA (2 hours for stop-loss) keeps issues from aging into formal disputes. The post-action layer that identity and edge-enforcement stacks do not cover.
See the Pre-Dispute Network/ 03Use cases
Cleaner approvals, fewer false positives.
Verified agents check out without getting flagged. When something does go wrong, the Pre-Dispute Network surfaces it merchant-to-operator first, so resolution happens before the card network sees a chargeback.
For merchantsSix primitives. Compose your own agent-trust layer.
/ AID
Agent Identity
A verifiable identity for AI agents — cryptographic, revocable, and reputation-aware.
/ AWA
Agent Wallet Auth
Delegated spending permissions with proof that a real principal authorized the agent.
/ TXG
Transaction Guard
Policy-based spend limits, per-merchant controls, and real-time transaction gating.
/ BOT
Bot vs. Agent
Tell useful autonomous agents apart from scrapers and malicious bots at the edge.
/ MVF
Merchant Verification
A verified "agent-friendly" signal for merchant endpoints that have completed onboarding.
/ LOG
Audit Logs
A signed, per-trace evidence record of agent actions — queryable through the Pre-Dispute Network for dispute resolution.
The problem
Most fraud systems were built for humans and retrofitted for bots. Agentic commerce needs something else entirely.
AI agents don't have credit histories. They don't match your velocity models. They look like bots to your WAF and like humans to your checkout. 3-D Secure fails them. Your chargeback policy can't locate them. Your evidence log has no row for them.
KnowYourAgent issues verifiable identity at the agent layer: cryptographic AID, a signed trace carrying the principal hash and spend scope, policy-bound transaction controls, and the Pre-Dispute Network that intercepts issues before chargebacks fire. Built for agents from the start — not bolted on.
/ 05Market context
The agent economy is already arriving at checkout.
1.3B
AI agents projected by 2028
IDC Research, sponsored by Microsoft (2025)
$3–5T
Agentic commerce by 2030
McKinsey & Company (2025)
$362B
Online payment fraud (2023–2028)
Juniper Research (2023)
“Agentic commerce is transforming the digital shopping experience. This shift will rival the impact that online shopping and mobile devices have had on commerce.”
“Sparky will be the primary digital vehicle for discovery, shopping, and managing everything from reorders to returns.”
/ 06FAQ
Common questions.
What is an AI shopping agent and why should merchants care?+
An AI shopping agent is autonomous software that browses, compares, and buys on behalf of a human customer. These agents are arriving at merchant checkouts today, and most fraud tools flag them as suspicious because they don't behave like human buyers — costing merchants legitimate revenue.
How does KnowYourAgent verify AI agents at checkout?+
Two calls at the checkout edge. The merchant calls /api/v1/verify with the agent's KYA ID and transaction context; KYA resolves the identity, the operator behind it, and the current trust level, and returns an accept / review / decline recommendation with reasons. Separately, the merchant opens /api/v1/checkout/sessions with the signed trace the operator issued upstream — that call validates the trace against a delegated wallet and enforces scope, spend cap, audience, and MCC before the card is touched.
What happens if a KYA-verified agent transaction goes wrong?+
The KYA Pre-Dispute Network runs as a side-channel between merchants and operators so problems surface before they escalate to chargebacks. If an agent exceeds its authorized scope or a customer disputes a purchase, an alert is opened against the X-KYA-Trace-ID with the signed evidence chain attached. Operators have 24 hours to respond (2 hours for stop-loss); configured merchant rules can auto-resolve matching alerts when they are opened.
How long does it take to integrate KnowYourAgent?+
Merchant onboarding is being packaged for a guided Shopify rollout, and custom API integrations fit alongside your existing checkout logic. Merchants start with /api/v1/verify for identity and recommendation, and add /api/v1/checkout/sessions when they want signed-trace enforcement at checkout. In both cases, KYA sits alongside your existing payment and fraud stack rather than replacing it.
How much does KnowYourAgent cost?+
Merchant onboarding currently starts with a guided pilot package, with launch pricing previewed on the pricing page. Operator pricing is separate and remains available for teams registering and managing agent fleets.
Does KnowYourAgent affect the checkout experience for human customers?+
No. KYA runs as a parallel verification call alongside your existing payment processing and only activates when an AI agent is detected. Human customers see no change to the checkout flow.