People & governance
Consent is read, never assumed.
Where a capability depends on a person's consent, the state is read from your consent system rather than assumed or cached here. An absent or unreadable state is treated as a refusal, which is the only safe default.
What connects
Your consent management platform, through its API.
What crosses the boundary
- Into the agent: the consent state relevant to the capability being used.
- Out of the agent: nothing.
- Never: writing a consent state, or recording one here.
The capabilities this exposes
- Read consent state
Setting it up
- 1Authorise a read-only connection.
- 2Map which capabilities depend on which consent categories.
- 3Confirm the behaviour on an unreadable state — refusal is the default and cannot be changed to permit.
Permissions
Read-only. We do not request the ability to record, alter or infer consent. Your consent system remains the only place a person's choice is recorded.
Jobs an agent can own once this is connected
- Use individual behavioural context only where consent permits
- Refuse a personalised message where consent was withdrawn
- Stop processing immediately when a state changes
Where it stops
The platform stores no consent state of its own — it reads yours, per use. If your consent system is unreachable, capabilities that depend on it refuse rather than proceeding on a cached answer. A withdrawal takes effect at the next read, which is immediate rather than on a schedule. Failing open is not a configuration option.
People & governance
Others in this category.
Give AI a job.
Build an agent, choose what it can access, define what it may do, and deploy it to work safely across your business.
No code required · Full SDKs and API · Stop any agent instantly · [email protected]