Generative AI Fundamentals Map

Map/Trust and governance/Wave 6

Agent identity and permissions

Who an agent is, who it acts for and what it may do.

capabilityEmerging Open on the map →

Facts

Lane: Trust and governance
Wave: 6 of 12 (build order, not a date)
Needs: 0 boxes all the way back through hard links
Unlocks: 27 boxes downstream through hard links
Leverage: 1.4 points · stands in the way of 9 of 53 stages · rank 8 of 57
Cluster: Trust plumbing and the agent economy

Evidence as of October 2026

Chan et al. propose agent IDs, binding agents to a legal person, and certification. A2A 1.0 added signed agent cards.

Around it

Needs · 0

Nothing beyond the detail list. It can start at any time.

Unlocks · 10

Agents that trade for investorscapability · Emerging
Agent-to-agent protocolscapability · Emerging
Agent-native organisationsstage · Emerging
Agent paymentscapability · Emerging
Audit and attributioncapability · Emerging
Reputation and credentialscapability · Speculative
Personal knowledge portabilitycapability · Speculative
Always-on background agentsstage · Emerging
Drop-in remote workerstage · Speculative · helps, not required
Model weight securitycapability · Emerging · helps, not required

Openings researchers' judgement, October 2026 scouting

Startup room 57/100 · rank 16 of 124 Research room 47/100 · rank 75 of 124

Crowding5 Lab overlap5 Small-team fit2.5 Buyer today4 Research openness3

Depth: deep
Players: Keycard ($38m); Oasis (Cyera ~$1bn); Okta/Auth0; Microsoft Entra Agent ID; WorkOS ($2bn); 20+ funded NHI vendors
Wedge: Protocol-edge components: delegation chains (user to agent to sub-agent to tool), cross-registry verification and machine-speed revocation, as reference implementations. startup