In reviewPDFWorksheetImplementationTeams and organizations

Review an AI pilot against evidence, quality, safety, adoption, and operational fit before deciding what happens next.

A structured scorecard for turning pilot observations into a clear continue, revise, expand, pause, or stop decision.

Preview · About 20 minutes

Resource download

Download the resource.

Use the downloadable preview now, then return to the written instructions and responsible-use notes below.

At a glance

The downloadable resource and its complete written workflow are available on this page.

Compare the pilot result with the baseline, target, and evidence collected.

Record errors, rework, exceptions, privacy concerns, and review reliability.

Name the decision, owner, conditions, and next review date.

Instructions

Use these written steps alongside the downloaded resource.

Return to the agreed outcome
Write the original pilot question, baseline, success measure, test group, and review period.
- Use observed evidence rather than impressions alone.
- Note any changes to the workflow or scope during the pilot.
Score each decision area
Review evidence, output quality, safety, human-review effort, adoption, and operational fit.
- Add a short evidence note beside every score.
- Record exceptions and failed cases, not only successful examples.
Review risks and stop conditions
Check whether any privacy, security, quality, safety, equity, or workflow stop condition occurred.
- Escalate unresolved concerns to the accountable owner.
- Do not average away a serious safety or privacy issue.
Make and record the decision
Choose continue, revise, expand, pause, or stop, then record the owner and conditions.
- Name the next action and review date.
- Keep the scorecard with the pilot evidence and decision record.

Responsible use

Do not place patient information, credentials, or confidential records in this scorecard.
A score does not override organizational approval, qualified review, or a triggered stop condition.