Best AI Prompts for Nurses

AI summary

Seven AI prompts for bedside and charge nurses that scaffold documentation and communication without crossing into independent clinical decision-making: SBAR hand-offs, patient education translation, care plan refresh, family conversation prep, med-pass pre-check, charting catch-up, and charge briefs. Every output is the nurse’s; AI is the structural co-pilot.

Nurses do the work that makes a unit run, mostly under time pressure no one outside healthcare understands. The seven prompts below take the parts of nursing where AI compounds (documentation drafting, education translation, hand-off structure, med-pass safety checks, charge briefs) and structure them so the nurse spends time on the bedside work that AI cannot do. This is the nurse slice of the AI Prompt Library, paired with a connector callout for the practice tools nurses use. For related medical-side prompts see Best AI Prompts for Doctors.

Why do most AI nurse-AI workflows produce documentation the chart audit would reject and slow down a shift?

The default nurse-AI risk is the same as the default physician risk, with an extra wrinkle: nurses work in higher-volume, faster-turnaround conditions where the temptation to let AI fill in clinical detail is even stronger. The result, if unchecked, is documentation that does not match the patient, hand-offs that miss the thing that matters, education materials that confuse rather than empower. Every prompt below is structured to keep the clinical content under the nurse’s control.

Use AI for the structural prep, the translation work, the safety check that catches what a tired brain might miss. Always review before any output reaches the chart, the patient, or the next shift. If you draft any patient-facing content with AI, run it through How to Edit AI Out of Your Writing before sharing. When a prompt becomes a shift-by-shift habit, graduate it using the Prompt-to-Workflow Ladder.

What are the seven for nurses prompts?

Prompt 1

Hand-Off Drafter (SBAR)

Most shift hand-offs are rushed and lose detail in the gap. This prompt structures the hand-off note from your shorthand so the next nurse picks up clean.

Hand-off for patient (de-identified):

ROOM: [GENERAL DESCRIPTOR, NOT IDENTIFIER]
SITUATION (presenting issue / admit reason): [BRIEF]
BACKGROUND (relevant history, allergies, code status): [BRIEF]
ASSESSMENT (current status, vitals trend, what changed this shift): [BRIEF]
RECOMMENDATIONS / PENDING (orders to follow up, things to watch, pending labs or imaging): [BRIEF]

Draft an SBAR hand-off note:

1. SITUATION: one sentence that names the patient's current clinical picture.
2. BACKGROUND: 2-3 bullets on what the next nurse needs from history.
3. ASSESSMENT: factual current status, including any change this shift.
4. RECOMMENDATIONS: what is pending, what to watch for, what to escalate.
5. THE ONE THING NOT TO MISS: the single most important detail the next shift should know.
6. FAMILY OR PSYCHOSOCIAL NOTES: anything about communication style, family dynamics, decisional capacity worth flagging.

Do not invent clinical details. Use SBAR convention consistently. Preserve clinical uncertainty ("reported," "observed," "appears") rather than stating things as fact.

When to use: Last 10 minutes of your shift, before report. · Best model: Claude (most disciplined about preserving clinical uncertainty).

Prompt 2

Patient Education Translator

Discharge instructions written in clinical language do not survive the parking lot. This prompt translates them into language the patient and family can actually act on.

Here are the discharge or patient-education instructions (de-identified):

[PASTE THE INSTRUCTIONS or describe what they cover]

Patient context (literacy level, primary language, family support): [BRIEF]
The action the patient needs to take: [SPECIFIC]

Translate into plain language:

1. WHAT HAPPENED: 2 sentences on the visit or admission in the patient's words, not clinical language.
2. WHAT TO DO AT HOME: numbered list of actions, each one a single specific step ("take this pill twice a day with food" not "as directed").
3. WHAT TO WATCH FOR: warning signs that mean call the clinic, written in plain English, with the specific phone number.
4. WHAT TO WATCH FOR THAT MEANS 911: separate list, never blended with the call-the-clinic list.
5. WHO TO CALL FOR WHAT: phone numbers labeled clearly (the on-call line, the prescription refill line, the follow-up scheduling line).
6. FOLLOW-UP: appointment dates and what each is for.

Keep at a 6th-grade reading level. Avoid medical jargon, or define it when unavoidable. Do not invent dosages or schedules. If the original instructions are ambiguous, flag that I should clarify with the provider before sending.

When to use: Before discharge, ideally before the patient is sitting in the wheelchair waiting. · Best model: Claude. Discipline about preserving clinical uncertainty matters.

Prompt 3

Care Plan Refresher

Most nursing care plans get written once and never revisited. This prompt updates the plan from your shift observations so it actually reflects current state.

Patient's existing care plan (de-identified):

NURSING DIAGNOSES: [LIST]
GOALS: [LIST]
INTERVENTIONS: [LIST]

What I observed this shift that suggests the plan needs updating:

[OBSERVATIONS: tolerance changes, new symptoms, family input, response to interventions]

Refresh the plan:

1. WHICH DIAGNOSIS NO LONGER FITS based on today's data. Suggest removal or revision.
2. WHICH NEW DIAGNOSIS the observations support. Drafted in NANDA language with the related-to and as-evidenced-by elements.
3. WHICH GOAL needs revision: from no-longer-relevant to current.
4. WHICH INTERVENTION needs adjustment: dose, frequency, approach.
5. ONE NEW INTERVENTION to consider, with rationale.
6. WHAT TO MEASURE in the next shift to know if the changes are working.

Do not invent observations. Build only from what I provided. Preserve clinical uncertainty.

When to use: End of each shift on a patient with active care issues. · Best model: Claude. The NANDA-language discipline benefits from a model careful with clinical terminology.

Prompt 4

Difficult Family Conversation Prep

You have to deliver news to a family that they will not want to hear. The first conversation shapes everything. This prompt prepares the framing.

I need to have a difficult conversation with a patient's family (de-identified):

THE NEWS: [WHAT NEEDS TO BE COMMUNICATED]
WHAT THE FAMILY ALREADY UNDERSTANDS: [BRIEF]
WHAT THE FAMILY HAS EXPRESSED CONCERN ABOUT: [BRIEF]
MY ROLE (am I delivering the news or supporting the provider who is): [ROLE]
CULTURAL OR LANGUAGE CONSIDERATIONS: [IF KNOWN]
WHO ELSE IS IN THE CONVERSATION (social work, chaplain, attending): [TEAM]

Draft a framework for the conversation:

1. THE OPENING: how to set up the conversation so the family is not blindsided.
2. THE LISTENING FIRST: what to ask before delivering so I know where they are starting from.
3. THE NEWS: how to deliver it factually without softening to the point of misinforming.
4. THE PAUSE: where to stop and let them respond.
5. THE QUESTIONS THEY ARE LIKELY TO ASK and how to handle the ones I cannot answer (defer to provider, defer to social work, name explicitly).
6. THE NEXT STEP: what they leave the conversation with (a contact, a follow-up time, a written summary).
7. THE CULTURAL OR LINGUISTIC ADJUSTMENTS to make in delivery.

Tone: warm, direct, never falsely reassuring.

When to use: Before the conversation, when there is time to think it through. · Best model: Claude. Tone discipline matters more than speed.

Prompt 5

Med-Pass Pre-Check

Med errors cluster around drug-drug interactions, allergies, and look-alike-sound-alike confusion. This prompt is a structured second look before pass.

Med list I am about to pass (de-identified):

[LIST OF MEDS WITH DOSE AND ROUTE]

Patient context: [ALLERGIES, KIDNEY OR LIVER FUNCTION IF RELEVANT, RECENT CHANGES]
Key labs that affect this pass (if recent): [LIST]

Pre-check the pass:

1. POTENTIAL INTERACTIONS: any drug-drug interactions in this list that I should be aware of, with the clinical significance.
2. ALLERGY CONFLICTS: any med that conflicts with the documented allergies.
3. LOOK-ALIKE-SOUND-ALIKE RISKS: any pair of meds in this list (or with common pharmacy errors) that I should double-check.
4. RENAL OR HEPATIC DOSING: any med where the dose may need adjustment given the function indicated.
5. TIMING CONSIDERATIONS: any med where timing relative to food, other meds, or labs matters.
6. PATIENT-SPECIFIC CHECK: anything about this patient's situation that would prompt me to verify with the pharmacist or provider before giving.

This is a CHECK, not a substitute for the pharmacist or provider. Flag anything uncertain rather than reassuring me. Do not invent contraindications.

When to use: Before any med pass that includes a med you have not given recently or a med change since last shift. · Best model: Claude. The discipline about flagging rather than reassuring matters here.

Prompt 6

Charting Catch-Up Drafter

End of shift, you have an hour of charting and 20 minutes. This prompt drafts the structure of pending notes from your shorthand so you fill in the substance, not the format.

Pending charting tasks (de-identified):

[FOR EACH: type of note (assessment, intervention, response, education), brief shorthand of what to document]

For each, draft the note structure:

1. NOTE TYPE: identify which charting category it belongs to.
2. KEY ELEMENTS the note must capture (objective vs subjective, time of intervention, response, follow-up).
3. THE SKELETON: a brief outline with placeholders for the specific clinical detail I need to add.
4. THE PHRASING that would be appropriate for the modality (assessment language, intervention language, education-response language).
5. THE COMMON OMISSION for this type of note that I should make sure I include.

Do not write the full note for me. Give me the structure; I fill in the clinical detail. If my shorthand is ambiguous, flag what to clarify before I document.

When to use: Last hour of shift, when charting is piling up. · Best model: Claude. The discipline about not writing the full note matters; I am the one charting.

Prompt 7

Charge-Nurse Daily Brief

Charge nurses run the shift on instinct and a clipboard. This prompt structures the morning brief so the team starts on the same page.

Census and status:

[FOR EACH UNIT/AREA: bed count, acuity descriptor, expected admits, expected discharges, expected transfers]

Staffing:

[NUMBERS: nurses, techs, on-call, any holes in coverage]

Known issues for the shift:

[OPERATIONAL: equipment, isolation precautions, pending IT issues, family situations needing extra attention]

Draft a 5-minute charge brief:

1. THE CENSUS PICTURE in one sentence.
2. THE PRESSURE POINT: the unit or shift segment most likely to be the bottleneck.
3. THE STAFFING REALITY: who is covering what, where the gaps are, how we will handle them.
4. THE WATCH-LIST PATIENTS: 2-3 patients (de-identified) whose status changes would shape the shift.
5. THE ONE OPERATIONAL ITEM: equipment, IT, family, that the team should know about.
6. THE ASK: what I am asking the team to bring me without me having to chase it.

Keep under 200 words. Tone: calm, organized, never alarmist. The brief sets the tone for the shift.

When to use: Start of shift, after the hand-off but before the team scatters. · Best model: Claude. Tone calibration matters; the brief is a leadership artifact.

These work across Claude, ChatGPT, Gemini, and Grok. Claude is the strongest default for nursing documentation because of its discipline about preserving clinical uncertainty and not fabricating clinical detail. For HIPAA compliance, the same rules as for physician use apply: paid plans with a signed Business Associate Agreement (BAA) where PHI is involved, and verification with your institution’s policy on which AI tools are approved. De-identify all inputs by default.

What is the worst thing you can do with AI for nurses?

Three patterns will sink nurse-AI workflows fastest, and all three have patient-safety weight.

  • Pasting identifiable PHI into a free-tier AI tool. Free tiers may train on inputs and may retain content. This is a HIPAA violation. Use only paid plans with verified data-handling and, where required, a BAA. The prompts above are designed to use general descriptors instead of identifying details.
  • Letting AI auto-draft chart documentation without your clinical review. AI fabricates clinical content confidently. Every chart entry is your responsibility. Use AI to structure what YOU observed; never let AI fill in observations you did not make.
  • Using AI to interpret labs, vitals, or images. AI is not a clinical decision-support system. The med-pass pre-check prompt is explicitly a check, not a substitute for the pharmacist or provider. Use AI to surface things to verify; verify with the licensed clinician.

What if you want to take this further?

Each prompt above takes inputs you paste in. The next move is connecting AI to the non-PHI workflow tools where shift coordination happens.

Connectors are now standard

Claude, ChatGPT, and Grok all support connectors that let your AI read live data from your work tools (Gmail, Notion, GitHub, Asana, HubSpot, Stripe, and many more) instead of relying on you to paste context. For nurses this means the AI can read appropriate non-PHI workflow tools (your team calendar, your unit scheduling tool, your CE tracker) while staying out of direct EHR access where institutional policy and HIPAA preclude it.

For nurses, the connectors worth pairing with these prompts:

  • Calendar connector — for shift-prep workflows that need to reference timing without pulling in chart content.
  • Notion connector — if your unit maintains non-PHI protocols, education materials, or shift-summary templates in Notion, AI reads them for consistency.
  • Google Drive connector — for de-identified case discussion, in-service prep, or research summary work.
  • Gmail or Outlook connector — for non-PHI team communication and CE tracking, AI references prior threads.
  • DO NOT use — consumer AI connectors with your actual EHR (Epic, Cerner, Meditech) unless your institution has explicitly approved the integration with a signed BAA and audited data flow.

What are common questions about AI for nurses?

Is AI use in nursing practice HIPAA-compliant?

It depends on the tool, the plan, and the data. Free-tier consumer AI is not HIPAA-compliant. Paid tiers with no-training and no-retention terms may be appropriate for non-PHI work and de-identified case discussion. For PHI handling, you need a Business Associate Agreement. Claude Enterprise, ChatGPT Enterprise, and nursing-specific AI tools may offer BAAs. Verify with your institution’s compliance office.

Will AI replace nurses?

No. The work that compresses (documentation, hand-off structure, patient education translation, charge-brief drafting) gets faster. The work that does not compress (bedside assessment, the read of the patient, family communication, code response, the moment-to-moment judgment of a high-acuity shift) is more important than ever. Nurses who use AI for the structural work and spend their saved time on actual patient care become more valuable, not less.

Which AI tool is best for nurses?

For documentation drafting on de-identified content, Claude Pro is most disciplined. For PHI handling, you need a HIPAA-compliant tool with a BAA: Claude Enterprise, ChatGPT Enterprise, or nursing-specific tools (Iris Telehealth, Hippocratic AI for certain niches). Verify what your institution has approved.

Can AI write my SBAR hand-offs?

AI can structure the SBAR template from your shorthand, which is what the Hand-Off Drafter prompt above does. AI should not generate clinical content from a blank prompt. Always review and verify every detail before report. The hand-off is your professional record.

Should I tell patients I use AI?

If AI affects a clinical decision or generates patient-facing content, disclose. If AI assists with workflow only, most jurisdictions do not require disclosure. Check your institution’s policy. Many institutions are developing AI-disclosure standards now; ask the patient experience or compliance team for the current guidance.

How do I avoid AI fabrication in clinical notes?

Never ask AI to fill in clinical detail you did not provide. Use AI to structure what YOU observed and reasoned; never let AI generate clinical content from a blank prompt. Every prompt above is built to keep AI in the structural role and to require nurse review. If AI generates a clinical detail you did not provide as input, do not use it.

How long does it take to build the nurse-AI loop?

Three weeks. Start with the Patient Education Translator and the Hand-Off Drafter on de-identified inputs first. Add the Med-Pass Pre-Check when you have a complex pass. Most nurses settle into 3-4 of the seven prompts as part of their shift rhythm within a month.

🎯

The AI Prompt Library · $39

Nursing workflows, prompt-paved.

Soon to be 1000+ prompts in Notion organized by use case. The full nursing section includes everything above plus prompts for in-service drafting, preceptor feedback, certification exam study planning, and union-grievance writing. Plus prompts for every other field. Lifetime access.

Get the Library →

Get Smarter About AI Every Morning

Free daily newsletter. Built for people who want to use AI well, not chase every model.

Free forever. Unsubscribe anytime.

Sources to read next?

You might also like

Two ways to go further

The AI Prompt Library

1,000+ ready-to-use prompts for Claude, ChatGPT, and Gemini. Stop staring at a blank box.

Get it for $39 →

2-Hour Live AI Crash Course

A private, beginner-friendly session across Claude, ChatGPT, Gemini, and the wider landscape.

Book for $125 →

Discover more from Beginners in AI

Subscribe now to keep reading and get access to the full archive.

Continue reading