Ready for agents?
Test one workflow—not an organisation, platform or agent—against the conditions it needs to operate. Leave with a clear Proceed, Prepare or Stop decision and the specific work still unresolved.

Decision
Output
Method Boundary & Connections
Start here when
Does not decide
Reassess when
Connection to IDAO
Connection to Agent Authority
Readiness before authority.
This framework decides whether the workflow has viable operating conditions. It does not decide how independently an agent may act.
Enter or update the responsible IDAO stage with the evidence recorded.
Turn missing conditions into work at the appropriate IDAO stage, then repeat the test.
Do not enter IDAO delivery for this scope; redefine it, resolve the blocker or stop.
For any selected consequential handover, set how independently it may act.
Evidence, not optimism.
Choose the statement that best matches current evidence. “Ready” must be demonstrable. One Stop condition stops the current scope; any Prepare condition names work to complete.
Name the bounded workflow, trigger, start and end point, business area and material exclusions. This stays in page memory and is included only in the local results PDF or optional print.
- 01
Workflow stability
Is the current workflow sufficiently understood and stable to specify?
- 02
Data & tool access
Can the work reach the right data and tools under enforceable permissions?
- 03
Observability
Can operators see what happened, why it happened and whether it worked?
- 04
Fallback & recovery
Can service continue safely when the agent or a dependency is unavailable?
- 05
Exceptions & boundaries
Are unusual, ambiguous and prohibited cases identifiable before harm occurs?
- 06
Operating economics
Does the workflow remain worthwhile after control, exception and run costs?
Current decision · 0/6 answered
Prepare
Complete the six conditions before making the decision.
Resolve the named conditions below. Reassess before committing the workflow to agent delivery.
Answer all six conditions to generate a complete local result.
Resolve the conditions, not the score.
The output is an operating-condition register. It records the gap, evidence required, accountable owner and reassessment date.
Map the current workflow, name its owner, quantify baseline performance and agree which variants are in scope.
Confirm source authority, identity, least-privilege permissions, tool contracts, data quality and retention.
Define case-level traceability, outcome measures, alerts, review cadence and a named monitoring owner.
Specify the fallback trigger, route, capacity, recovery point, reconciliation method and rehearsal evidence.
Build the exception inventory, prohibited-case rules, escalation route, service level and accountable resolver.
Evidence demand, cost-to-serve, exception effort, control overhead, adoption and the threshold for stopping.
Proprietary method. Public evidence.
External source review: 10 September 2026. Applicability depends on jurisdiction, sector, system classification and intended use.
- NIST AI Risk Management Framework 1.0 (January 2023)
Govern, Map, Measure and Manage functions informed the evidence and monitoring questions.
- NIST AI 600-1, Generative AI Profile (July 2024)
Risk identification, measurement, incident handling and third-party dependency considerations informed the operating-condition prompts.
- EU AI Act, Regulation (EU) 2024/1689 (13 June 2024)
Human oversight, logging, accuracy, robustness and cybersecurity obligations informed the control questions where applicable.
Prepare what is missing. Then earn the right to operate.
A readiness decision defines whether the workflow should enter delivery. IDAO builds and proves the capability; Agent Authority governs each live handover.