Request a fit review

Decision tool · AI integration

Is this workflow ready for an AI pilot?

Assess eight foundations that determine whether implementation can produce evidence or will only hide uncertainty. The result is immediate and ungated.

0 of 8 assessed

Q01How clearly is the workflow defined?

Consider trigger, owner, steps, output, and completion.

Q02How representative is the available data?

A pilot needs normal and difficult cases, not only ideal examples.

Q03Can the required systems be accessed safely?

Include APIs, exports, test environments, and accountable owners.

Q04Can good output be judged?

The test can combine rubrics, schemas, business rules, and review.

Q05Are common errors and exceptions understood?

Production design depends on known failure categories.

Q06Where will human approval remain?

High-consequence or uncertain cases need an explicit owner.

Q07Have sensitive-data boundaries been mapped?

Include providers, storage, logs, retention, and permissions.

Q08Is there a measurable current baseline?

Examples include handling time, error, queue size, or review effort.

Scoring method

The score rewards observable foundations, not enthusiasm.

Each answer contributes zero, two, or three points. A high band means the workflow appears suitable for a controlled pilot; it does not guarantee model quality or production approval. Any low-scoring foundation is shown as a blocker because one missing control can matter more than the average.

Technical fit review

Use the result to choose the next responsible step.

Share the system, workflow, or delivery risk you need to resolve. The first review focuses on fit and a practical next step.

Request a Technical Fit Review