Service
Read the case, handle it or hand it over.
Agents can read cases assigned to them, open new ones and escalate with everything gathered attached. Cases marked sensitive are excluded from agent processing by default, and that exclusion lives on the case rather than on the agent.
What connects
Your customer service or ticketing system, through its API.
What crosses the boundary
- Into the agent: type, status, owner, timeline and history of cases in its scope.
- Out of the agent: new cases, escalations, and notes attributed to the agent.
- Never: cases marked sensitive, or cases outside the agent's assignment.
The capabilities this exposes
- Read a case
- Create a case
- Escalate a case
Setting it up
- 1Authorise the connection.
- 2Confirm which case types agents may handle.
- 3Confirm which are marked sensitive and excluded.
- 4Grant capabilities per agent.
Permissions
Read within scope, create, and escalate. We do not request the ability to close a case, delete one, or alter its history. An agent that could resolve its own case could hide a problem by closing it.
Jobs an agent can own once this is connected
- Answer routine cases from approved knowledge
- Escalate anything outside what it can verify
- Assemble a service history for an account review
Where it stops
An agent cannot close a case, and that is deliberate. It can propose a resolution and hand over. Sensitive cases are invisible to agents by default and making them visible is an explicit decision per case type, not a global switch.
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]