Rapid Launch: one fixed-scope AI workflow delivered in 24 hours See the package →

Free readiness scorecard · Published August 11, 2026

AI automation readiness assessment

Short answer: a workflow is ready for an AI automation pilot when its steps and owner are clear, value is measurable, data and systems are accessible, permissions and human review are bounded, acceptance tests exist, and someone owns operations after launch. Score one workflow below to find the next useful action.

12 evidence checks

Score one workflow

Yes = 2 · Partly = 1 · No = 0

1.The workflow is recurring and frequent enough to justify automation.
2.Its trigger, normal steps, output, and common exceptions can be written down.
3.A named business owner can approve scope and resolve process questions.
4.The current time, cost, delay, error rate, or missed revenue has a measurable baseline.
5.Representative normal, edge, and failure-case inputs are available for testing.
6.The required source-of-truth fields are sufficiently complete and consistent.
7.The necessary systems provide usable APIs, webhooks, exports, or another stable interface.
8.Data sensitivity, allowed users, access scopes, and retention constraints are known.
9.Actions the system may take—and must never take—are explicitly defined.
10.Human approval and escalation rules are defined for uncertainty or consequential actions.
11.Objective acceptance tests can determine whether the workflow is safe and useful.
12.Someone will own monitoring, exceptions, vendor changes, rollback, and maintenance after launch.

Privacy: scoring runs locally in your browser. Your selections are not transmitted or stored.

Interpretation

What the score means

19–24

Define a bounded pilot

Confirm the missing evidence, write acceptance tests, and start with limited permissions, representative cases, and a rollback path.

12–18

Authorize discovery first

Map the workflow, validate integration access, gather test cases, and quantify the baseline before committing to production scope.

0–11

Repair the foundation

Clarify ownership and process, improve source data, or choose a more stable and measurable workflow to assess.

A score is a prioritization aid, not a risk certification. Any “No” involving sensitive data, consequential decisions, security, legal duties, or human oversight requires specific review even when the total is high.

Primary guidance

Why the assessment uses these checks

  • NIST AI RMF Core connects an initial go/no-go decision with documented purpose, business value, scope, data, human oversight, evaluation, monitoring, and ownership.
  • Australia’s National AI Centre screening questions recommend checking control needs before committing resources and treating early screening as distinct from a full risk assessment.
  • UK AI procurement guidance emphasizes problem-first requirements, data assessment, testing, lifecycle planning, transparency, and avoiding lock-in.

FAQ

AI automation readiness questions

How do I know if a process is ready for AI automation?

A process is a strong candidate when it is recurring, clearly owned, measurable, supported by accessible data and systems, bounded by explicit permissions, testable with representative examples, and assigned to someone after launch.

Which business process should we automate first?

Compare one workflow at a time. Start with the highest-readiness process that is frequent, painful, measurable, reversible, and low enough in consequence to pilot safely. Avoid choosing only by theoretical savings.

Does a low score mean we should abandon automation?

No. It usually means the next investment is discovery or process cleanup rather than development. Resolve unclear ownership, inconsistent steps, missing data, unavailable access, or absent controls, then score the workflow again.

Does a high readiness score replace a risk or compliance assessment?

No. This is an early project screen, not legal, security, privacy, compliance, or procurement approval. Higher-impact uses need deeper review appropriate to their context and jurisdiction.

Does CTB receive the answers entered in this assessment?

No. Scoring runs in your browser. The selections are not transmitted or saved by CTB.

Start with one workflow

Found a pilot-ready workflow? Pressure-test it before building.

Share the workflow without secrets or live credentials. We’ll test the assumptions, controls, integration path, and smallest useful delivery scope.

Review my workflow