ConsentGraph

Web operators increasingly receive requests from browsers, scripts and delegated agents but lack a consistent way to distinguish declared automation, verify credentials and apply consent or access policy. ConsentGraph collects minimal client and server evidence, verifies any presented agent credential against approved registries and returns allow, challenge, limit, block or log decisions. The supplied research confirms enterprise agent-traffic measurement, a public verification interface using signed credentials and rapid growth in agentic traffic. Enterprise bot and identity products are observed market references, not fixed product pricing. Detection rates vary materially by source and segment, so traffic classification is a probabilistic candidate, not proof that a visitor is human, automated or acting for a named principal. Session evidence, presented credential, cryptographic verification, registry status, behavioral classification, policy decision, challenge outcome, application authorization and downstream action remain separate. A signed receipt proves only what the issuer recorded and signed; it does not prove the requester, delegation or behavior was truthful. The product can centralize policy and preserve decision evidence. It cannot guarantee identity, defeat every bot, replace application authorization or covertly fingerprint people beyond a documented lawful purpose.

A web platform security, fraud or trust leader receiving meaningful automated traffic and willing to publish an explicit agent-access policy.