Care Experience Lab · Prompt Library
AI Prompt Library for Healthcare Leaders
Copy-and-paste prompts to help leaders use generative AI safely and responsibly to improve communication, patient experience, leadership, quality, and operations.
Lab: Responsible AI in Healthcare · Type: Prompt Library · Audience: Healthcare Leaders · Difficulty: Intermediate · Version 1.0
Executive Summary
Generative AI can save leaders hours on drafting, summarizing, and analyzing — but only when used carefully. This library gives you ready-to-use prompts for the work leaders actually do, organized by task. Every prompt is copy-and-paste ready and paired with a note on when to use it and what a human must still review. Used well, these prompts speed up the busywork so you can spend more time with patients and teams.
Responsible AI Disclaimer
Read before using any prompt.
- Never paste protected health information (PHI) or other confidential data into AI tools that are not approved and covered by appropriate agreements at your organization.
- AI can be wrong. It can invent facts, statistics, and citations. Verify every factual claim against a trusted source.
- AI does not replace clinical or professional judgment. Use it for drafting and analysis support, not decisions.
- A qualified human must review and approve anything before it is published, sent, or acted on.
- Follow your organization’s AI policy and governance requirements.
How to Use This Prompt Library
- Find the task category you need below.
- Copy the prompt and replace the [bracketed] parts with your details — without PHI or confidential data.
- Paste it into an approved AI tool.
- Review and edit the output; verify any facts; apply the required human review.
- Treat AI output as a first draft, never a final product.
Communication Prompts
Draft a clear staff message
Help me write a short, clear message to my team about [topic/change]. Audience: [roles]. Tone: respectful and supportive. Keep it under [150] words, plain language, and end with what they should do next.
When to use: Announcing a change, update, or expectation to staff.
Human review required: Confirm accuracy of facts/dates; ensure tone fits your culture; check nothing confidential is included.
Improve a confusing message
Rewrite the following message to be clearer, warmer, and easier to act on. Keep the meaning the same. Use plain language and short sentences:
[paste your draft — no PHI]
When to use: You have a draft that feels unclear or too formal.
Human review required: Verify the meaning wasn’t changed; confirm any specifics are still correct.
Prepare difficult-conversation talking points
Help me prepare talking points for a sensitive conversation with [a staff member / a team] about [general topic]. I want to be honest, respectful, and solution-focused. Give me an opening, key points, and a constructive close.
When to use: Preparing for a hard but important conversation.
Human review required: Adapt to the real person and context; ensure fairness and compliance with HR policy; do not include identifying details in the prompt.
Patient Education Prompts
Create plain-language patient instructions
Turn the following instructions into plain language at about a 6th-grade reading level. Use short sentences, ‘you’, and a numbered list of steps. Add a ‘call us if…’ section:
[paste general, non-PHI instructions]
When to use: Simplifying instructions or handouts for patients.
Human review required: A clinician must verify clinical accuracy before patient use; confirm reading level and that no step was lost.
Generate teach-back questions
Give me 3 teach-back questions to confirm a patient understands [topic, e.g., how to take a new medication]. Make them open-ended and easy to answer in the patient’s own words.
When to use: Building teach-back into patient education.
Human review required: Confirm questions fit the actual instruction; clinician review for accuracy.
Health Literacy Prompts
Check and simplify reading level
Estimate the reading level of the text below and rewrite it to about a 6th-grade level without losing key meaning. List any medical terms you simplified:
[paste non-PHI text]
When to use: Reviewing materials for health literacy.
Human review required: Reading-level estimates are approximate; verify clinical meaning is intact; clinician sign-off for patient materials.
Build a jargon-to-plain-language list
Create a table of common terms for [service line/topic] with a plain-language alternative for each, suitable for patient communication.
When to use: Standardizing plain language across a team.
Human review required: Clinical reviewer should confirm the alternatives are accurate and not misleading.
CAHPS Analysis Prompts
Summarize CAHPS comment themes
Group the following de-identified patient comments into themes, rank them by frequency, and note any quick-win opportunities. Do not invent comments:
[paste de-identified comments only]
When to use: Making sense of open-ended HCAHPS comments.
Human review required: Ensure comments are fully de-identified; verify themes against the raw comments; AI may miss nuance.
Draft an action plan from a low domain
Our [domain, e.g., communication about medicines] CAHPS results are below target. Suggest 5 practical, evidence-informed actions a unit could take, focused on staff behaviors. Keep them realistic for a busy unit.
When to use: Turning a weak measure into an action plan.
Human review required: Confirm actions fit your setting and evidence; do not present AI suggestions as guaranteed to raise scores.
Home Health CAHPS Prompts
Improve in-home communication
Suggest practical ways a home health team can improve communication and continuity for patients between visits, focused on what clinicians say and do in the home.
When to use: Targeting HHCAHPS communication and care measures.
Human review required: Adapt to your agency’s workflows and regulations; verify feasibility with frontline staff.
Draft a caregiver teaching checklist
Create a simple checklist a home health clinician can use to teach a family caregiver how to safely manage [general task] between visits. Plain language, step by step.
When to use: Supporting caregivers in the home setting.
Human review required: Clinician must verify the steps are safe and complete for the specific situation; no PHI in the prompt.
Complaint Analysis Prompts
Identify root-cause themes in complaints
Review these de-identified complaints and group them by likely root cause (communication, access, environment, etc.). Suggest one improvement per theme. Do not fabricate details:
[paste de-identified complaints only]
When to use: Spotting patterns across multiple complaints.
Human review required: Ensure full de-identification; validate root causes with the team; AI groupings are a starting point, not conclusions.
Draft a service-recovery response
Help me draft a warm, accountable response to a patient concern about [general issue]. Acknowledge, apologize appropriately, explain the next step, and offer to follow up. Keep it sincere and under [150] words.
When to use: Responding to a patient or family concern.
Human review required: Verify facts and commitments before sending; ensure compliance with your complaint and privacy policies; never include PHI in the prompt.
Leader Rounding Prompts
Generate rounding questions for a focus area
Give me 5 open-ended leader rounding questions to use with [patients / nurses / physicians] focused on [topic, e.g., communication]. Make them genuine, not yes/no.
When to use: Refreshing your rounding questions.
Human review required: Adapt to your team and culture; pick a few rather than asking all.
Summarize rounding notes into actions
Turn these rounding notes into a short action list with owners and follow-up dates. Flag anything urgent:
[paste de-identified notes]
When to use: After a round, organizing what you heard.
Human review required: De-identify notes; confirm owners agree; you decide urgency, not the AI.
Staff Coaching Prompts
Prepare supportive coaching feedback
Help me frame constructive, specific feedback for a team member about [observable behavior, no names]. Use a supportive tone, focus on behavior not personality, and include a forward-looking next step.
When to use: Preparing for a coaching conversation.
Human review required: Tailor to the real person; ensure fairness and HR alignment; keep identifying details out of the prompt.
Create a skills-practice scenario
Write a short role-play scenario to help staff practice [skill, e.g., teach-back]. Include a setup, a patient line or two, and what ‘great’ looks like.
When to use: Training and team practice sessions.
Human review required: Confirm the scenario is realistic and respectful; clinician review if clinical detail is involved.
Psychological Safety Prompts
Plan a psychological safety discussion
Help me plan a 15-minute team discussion to strengthen psychological safety. Include 3 open questions and one commitment we can make together. Keep it blame-free.
When to use: Building psychological safety with your team.
Human review required: Adapt to your team; create genuine follow-through, not a one-time talk.
Reframe a blameful message
Rewrite the following message to be blame-free and learning-focused while still being honest about what needs to change:
[paste your draft — no names/PHI]
When to use: You have a message that may sound blaming.
Human review required: Confirm it still conveys accountability where needed; check tone fits the situation.
Meeting Summary Prompts
Summarize notes into decisions and actions
Summarize these meeting notes into: (1) key decisions, (2) action items with owners, (3) open questions. Be concise:
[paste de-identified notes]
When to use: Turning messy notes into a clean summary.
Human review required: De-identify notes; confirm decisions/owners are accurate before sharing.
Draft an agenda from objectives
Create a focused [30-minute] meeting agenda to achieve these objectives: [list objectives]. Include time estimates and a clear desired outcome for each item.
When to use: Planning an effective meeting.
Human review required: Adjust timing to reality; confirm objectives are the right ones.
Policy Drafting Prompts
Draft a first-pass policy outline
Draft a plain-language outline for a policy on [topic]. Include purpose, scope, principles, requirements, roles, and review. Mark anything that needs legal or compliance input.
When to use: Starting a new policy or procedure.
Human review required: Legal/compliance must review; verify against regulations; this is a draft, not an approved policy.
Simplify an existing policy
Rewrite this policy excerpt in plain language for staff, without changing its requirements. Flag anything ambiguous:
[paste non-confidential excerpt]
When to use: Making a dense policy understandable.
Human review required: Confirm meaning and requirements are unchanged; compliance review before use.
Workflow Redesign Prompts
Map and streamline a workflow
Here is a current workflow for [process]: [describe steps]. Identify likely bottlenecks, redundant steps, and 3 ideas to streamline it while protecting safety.
When to use: Looking for inefficiencies in a process.
Human review required: Validate with the people who do the work; never trade away safety for speed; confirm feasibility.
Design a pilot test
Help me design a small pilot to test [change]. Include a goal, who’s involved, success measures, a timeframe, and how we’ll gather feedback.
When to use: Planning a safe, measurable pilot.
Human review required: Confirm measures and feasibility; ensure appropriate approvals before piloting.
Prompt Writing Best Practices
- Give context: say who it’s for, the goal, and the tone you want.
- Be specific: length, format (list, table), and reading level.
- Use brackets for the details you’ll fill in — and never put PHI there.
- Ask for a draft, then refine: “make it shorter,” “warmer,” “simpler.”
- Request plain language when the audience is patients or families.
- Ask it to flag uncertainty and anything that needs human or legal review.
- Iterate — the second or third version is usually much better.
Common AI Mistakes
- Pasting PHI or confidential data into unapproved tools.
- Trusting facts, statistics, or citations without verifying them.
- Publishing or sending AI output without human review.
- Treating AI as a decision-maker instead of a drafting aid.
- Using vague prompts and accepting vague results.
- Forgetting to check tone and accuracy for your real audience.
- Ignoring your organization’s AI policy and governance.
References
- National Institute of Standards and Technology (NIST). AI Risk Management Framework (AI RMF 1.0), 2023.
- World Health Organization (WHO). Ethics and Governance of Artificial Intelligence for Health, 2021 (and 2024 guidance on large multi-modal models).
- Office of the National Coordinator for Health IT (HHS). Algorithm transparency provisions (HTI-1 Final Rule), 2024.
- U.S. Department of Health and Human Services / HHS HIPAA resources on protecting health information.
AI guidance and regulation evolve quickly; verify the current version and URL of each source, and follow your organization’s current AI policy, at time of use.
Related Care Experience Lab Resources
- Responsible AI Governance Toolkit for Healthcare Leaders
- Health Literacy & Plain Language Transformation Toolkit
- Explore the full Resource Library
Version 1.0 · Review date: [set at publication] · Care Experience Lab · Responsible AI in Healthcare. Always follow your organization’s AI policy. For general informational purposes; not legal, compliance, or clinical advice.
Healthcare Communication Clarity Assessment →
Use AI to support clearer patient communication — start by assessing clarity.