Readiness check

How well-defined and repeatable is the work you want agents to deliver?

1 = every request looks different and needs fresh judgement · 5 = the work follows the same repeatable steps every time, like a checklist.

How clearly can a person judge whether a piece of the work is done right?

1 = whether it's ‘done right’ is a judgement call · 5 = you can check it against clear pass/fail criteria.

How clear and agreed is the business outcome this initiative must deliver?

1 = we're not sure yet what success looks like · 5 = a named sponsor has agreed a specific, measurable target.

How realistic are expectations about what AI agents will achieve here?

1 = expectations are based on hype or assumption · 5 = expectations are grounded in evidence from a pilot or comparable case.

How ready is your way of working to handle output produced faster than people can review it?

1 = output would pile up faster than anyone can check it · 5 = we have people and time budgeted to review output at the pace it's produced.

How defined are the roles and accountabilities for agent-driven work?

1 = it's unclear who owns what · 5 = every step has a named owner from start to finish.

How well are quality gates and human sign-off built into the delivery flow?

1 = nothing stops work from shipping unchecked · 5 = a person signs off before anything goes live.

How clearly is accountability for agent decisions assigned to named people?

1 = no one is clearly accountable if something goes wrong · 5 = a named person is accountable — never the agent itself.

How ready are your delivery people to work with and supervise agents?

1 = our people haven't worked with agents before and aren't prepared to · 5 = our people know how to supervise and correct agent output.

How well is adoption owned — communication, training, and addressing concerns?

1 = no one is driving adoption or addressing concerns · 5 = someone owns a clear plan for training and buy-in.

How accessible, consistent, and current is the knowledge and data agents will rely on?

1 = our knowledge and data are scattered, outdated, or contradictory · 5 = our knowledge and data are centralised, current, and consistent.

How ready is your technical environment — access, repositories, security, tooling?

1 = agents wouldn't have safe, working access to our systems today · 5 = access, repositories, security, and tooling are all in place.

Is there a competent person who can validate the agents' output, and is the work not purely open-ended exploration?

This is a hard gate. Without a validator, or for purely open-ended work, agent-driven delivery is not yet the right path.

Is the knowledge and data agents need accessible, reasonably consistent, and current enough to work from?

This is a hard gate. A weak context foundation caps readiness until it is addressed.

How high is the business value of getting this initiative right?

Used to place the initiative on the readiness map (shown in the full report).

How complex and risky is this initiative?

Used to place the initiative on the readiness map (shown in the full report).