For clinical and operational teams

Plan a responsible healthcare AI pilot

Work through the evidence, the people responsible for oversight and the conditions for stopping before moving from a demonstration to a local evaluation.

Step 1

Define the task and evidence

Read A practical healthcare AI pilot and How to evaluate a healthcare AI claim. Identify one intended use and the current process it would be compared with.

Use the evidence worksheet to record what is known and what needs local investigation.

Step 2

Name the people and the boundaries

Review Human oversight in healthcare AI and Data foundations for healthcare AI.

Your next step: document who can review output, report problems and decide to pause the evaluation. Resolve privacy and data access questions before using operational data.

Step 3

Decide what would make you stop

Read When a healthcare AI system should be paused. Download the safety and pilot worksheet (PDF) to document ownership, outcomes and pause criteria.

A completed worksheet records a discussion. It does not establish that a system is safe, effective or approved for deployment.

Start with a concrete use case

For documentation tools, read what a clinical team should evaluate about AI scribes alongside the ambient scribe randomized trial evidence record.

For diagnostic tools, read why accuracy is only part of the decision. Keep a research result separate from evidence for routine care.

Bookmark this path and revisit your recorded assumptions before the next evaluation meeting.

Plan a responsible pilot

Define the task, human oversight and conditions for stopping.

Start the team path