Agentic AI

To do this in a trusted way, they need verifiable information about identity, permissions, capabilities, and provenance. Verifiable credentials and wallets are uniquely suited for this — enabling agents to carry and present trusted, machine-readable proofs across systems, without relying on centralized control or manual verification. A new set of verifiable credentials is emerging to support this:

– Agent identity & metadata credentials — who or what the agent represents including owner, purpose, and scope
– Delegation & authorization credentials — what the agent is allowed to do on behalf of a user or organization
– Agent capability credentials — what the agent can do (skills, services, APIs)
– AI model provenance credentials — which model is used
– Data provenance credentials — which data sources are involved

Together, these credentials enable agents to act in a way that is transparent, verifiable, and controllable — across organizations, systems, and ecosystems.

For users and regulators, this makes agent behavior understandable and governable. For developers and service providers, it enables automated, policy-driven interactions between agents and systems.