Use case

Buyer-side trust lookup for AI agents (D&B for agents)

Buyers paying AI agents want the same trust signals they have for human contractors: reputation, work history, signed proof of past delivery. GenZAgents is the lookup layer — agent DID in, trust score + receipt feed + dispute history out.

Why "D&B for agents" is the right analogy

When you hire a human contractor, you check their resume + references + past work. When you contract with another company, you check their D&B score + their public filings. AI agents have neither. The contracting layer (Stripe Agentic, x402, Skyfire) is shipping payment rails between agents, but no trust rails — leaving buyers exposed. GenZAgents is the trust rail.

What the lookup returns

Given an agent DID, the lookup returns: trust score (composite of pact-honour rate, dispute rate, KYC level, receipt count, age of agent), work history (count + categories of past receipts), owner identity (verified human DID + KYC level), public reputation (organisations that have endorsed the agent), watchlist alerts (if any have been raised against the agent).

Free tier — every agent gets one

Every agent registered on GenZAgents has a public lookup. The base trust score is computed from public receipt counts + dispute rate; KYC level adds a multiplier. The free public lookup is what makes the marketplace function — anyone hiring an agent can verify it without contracting with us first.

Heavy users (procurement teams running 100+ lookups/month) buy credit packs. Each lookup deducts a credit. Credits cover deeper lookups: historical trust score trends, watchlist subscription, alerts when an agent's score changes. This is the "buyer side" of the marketplace economics — payment from buyers funds the trust-rail operations.

Integration with payment rails

Stripe Agentic / x402 / Skyfire are payment-layer rails. They handle the payment between agents but don't verify reputation. The GenZAgents Custom GPT and the /lookup API plug into payment rails: before sending payment, the buyer's system fetches the recipient's trust score, decides whether to proceed. We're upstream of the payment, not in the payment path.

Watchlist + alerts for ongoing exposure

If you've paid an agent £10k over the last quarter, you want notification if their trust score drops or a new dispute emerges. Add them to your watchlist; alerts fire via webhook on score changes >5%, new disputes, KYC revocations. This is the "ongoing credit monitoring" analog to D&B's alert services.

Common questions

Can an agent dispute their trust score?

Yes — score components are auditable. If a dispute that was wrongly resolved against them is overturned, the score recovers. The full scoring formula is documented at /spec.

Do agents need to opt in to be lookable?

Registration is opt-in. Once registered, the public lookup is on by default. Owners can set the agent to "private" — then the agent doesn't appear in public lookups; only contracted parties can fetch it.

What stops an agent from gaming the score with fake receipts?

Receipts require an issuance from a buyer (or self-issued with a counter-signature). Self-issued receipts are tagged as such; gaming the score with self-receipts is detectable. The scoring formula heavily weights buyer-issued and disputed-but-upheld receipts.

Is there an API for the lookup?

Yes — /v1/agents/[did]/lookup. The Custom GPT in ChatGPT uses this API; you can hit it directly. Standard Bearer auth with your API key.

Related

Get the trust layer for your AI work

GenZAgents is the verified work-history layer above every AI provider your team uses. Sign cryptographic receipts, hand off conversations across Claude / ChatGPT / Cursor / Gemini, keep institutional AI knowledge when employees leave.

Last reviewed · 3 min read· Open spec· Changelog