Deployment readiness
The week-one owner map and a readiness score out of 100. See every connected system and agent, who owns its data, model risk and audit, and what still stands between the deployment and production.
- Find every system the agents touch
- Confirm a data, model-risk, audit and operational owner for each
- Read the readiness score and close its biggest gaps
Open Settings › Readiness (/readiness). It needs admin.write.
The owner map
The page lists what is connected, found from the platform's own records:
- Systems: every data source (type, status, records) and every MCP server.
- Agents: every registered agent, with its risk tier and whether it has an active owner.
For each system you confirm four owners: data owner, model-risk owner, audit owner and operational owner. For the instance as a whole you confirm an audit owner and a model-risk owner.
Discovery suggests candidates from the data: for a source, the teams that work its tickets, most tickets first; for an agent, its registry owners and oversight roles. Click a suggestion to use it. A suggestion never counts as an owner until someone confirms it.
The readiness score
The score is out of 100, from nine weighted checks:
| Check | Points | Full marks when |
|---|---|---|
| Systems connected | 15 | Every connected system is healthy |
| Sources synced | 10 | Every source synced in the last 7 days |
| Data owners | 15 | Every system has a confirmed data owner |
| Agent owners | 10 | Every agent has an active owner |
| Model-risk ownership | 10 | The instance has a model-risk owner, or every agent has one |
| Audit owner | 5 | The instance has an audit owner |
| Knowledge coverage | 15 | Every self-service question was answered (half marks while records are indexed but nobody has asked yet) |
| Agents governed | 10 | Every agent is under a guardrail or a budget of its own |
| Agents exercised | 10 | Every agent ran through the gateway in the last 90 days |
80 and up reads ready, 60 to 79 nearly, below 60 not ready. Each check says what it found and what closes it, and the three checks with the most points still to gain are shown as the blockers.
API
GET /api/admin/readiness: the map and the score.PUT /api/admin/readiness/ownerswith{"system": "source:<id>" | "mcp:<name>" | "agent:<key>" | "instance", "roles": {"data_owner": "maria@acme.example"}}. An empty value clears that role. Each change writesfde.owner.assignwith before and after.