Best AI Prompts for Vets

AI summary

Seven AI prompts for general-practice, specialty, and emergency veterinarians: SOAP notes, end-of-life conversations, estimate walkthroughs, discharge instructions, behavior plans, specialist referrals, review responses. Built to scaffold practice without crossing into clinical decision territory.

Prompt 1

SOAP Note Drafter

Most vet SOAP notes get written tired between appointments. This prompt structures the note from your shorthand.

Patient (de-identified):

SPECIES, BREED, AGE: [BRIEF]
PRESENTING COMPLAINT: [WHAT THE OWNER SAID]
MY EXAM FINDINGS shorthand: [BRIEF]
DIFFERENTIAL diagnoses: [LIST]
DIAGNOSTICS run or recommended: [BRIEF]
TREATMENT plan: [BRIEF]
CLIENT EDUCATION provided: [BRIEF]

Draft a SOAP note:

1. S: subjective from owner, in their language.
2. O: objective findings, with vitals and notable exam elements.
3. A: assessment with differential ranked.
4. P: plan with diagnostics, treatment, follow-up.
5. CLIENT EDUCATION noted.
6. NEXT VISIT cadence.

Preserve clinical uncertainty ("suspect," "consistent with," "cannot rule out"). Do not invent findings.

When to use: Same day as the visit. · Best model: Claude.

Prompt 2

End-of-Life Conversation Prep

Euthanasia conversations are the hardest part of vet practice. This prompt prepares the framing.

Patient context (de-identified):

SPECIES, AGE, DIAGNOSIS: [BRIEF]
OWNER'S RELATIONSHIP to the animal (long-time, recent, complex): [BRIEF]
WHAT THE OWNER HAS BEEN TOLD already: [BRIEF]
QUALITY OF LIFE indicators currently: [BRIEF]
MY CLINICAL READ on timing: [BRIEF]

Draft a conversation framework:

1. OPENING that meets the owner where they are emotionally.
2. THE QUALITY OF LIFE conversation in their language, not vet language.
3. THE TIMING question without imposing timing.
4. THE OPTIONS for the visit itself (home, clinic, who can be present).
5. THE LOGISTICS handled with dignity.
6. THE AFTERCARE conversation.
7. WHAT I will offer them when the visit is done.

Tone: warm, direct, never rushed. Do not minimize their grief.

When to use: Before the conversation. · Best model: Claude. Tone discipline matters.

Prompt 3

Client Estimate Walkthrough

Most cost conversations in vet practice go badly. This prompt drafts the walkthrough.

Estimate context:

DIAGNOSTIC or treatment recommended: [BRIEF]
ESTIMATED COST range: [LOW-HIGH]
WHAT IS INCLUDED: [LINE ITEMS]
WHAT MAY BE ADDED if findings change: [BRIEF]
ALTERNATIVES if budget is a constraint: [BRIEF]
MY CLINICAL READ on what is necessary vs nice-to-have: [BRIEF]

Draft the walkthrough:

1. THE OPENING that acknowledges money matters without apologizing.
2. THE ESTIMATE walked through line by line.
3. THE WHY for each line tied to the patient.
4. THE ALTERNATIVES with consequences spelled out.
5. THE WALK-AWAY: what happens if they decline entirely.
6. THE PAYMENT options the practice can offer.
7. THE QUESTION I leave with them.

Do not pressure. Do not over-promise outcomes.

When to use: Before presenting the estimate. · Best model: Claude. Tone discipline matters.

Prompt 4

Discharge Instructions

Most discharge instructions get skimmed. This prompt drafts them for clarity.

Discharge context:

SPECIES, DIAGNOSIS, PROCEDURE: [BRIEF]
MEDICATIONS prescribed: [LIST WITH DOSING]
ACTIVITY RESTRICTIONS: [BRIEF]
DIET CHANGES: [BRIEF]
WHEN TO CALL US (warning signs): [LIST]
FOLLOW-UP date: [SPECIFIC]
OWNER LITERACY: [BRIEF]

Draft the discharge:

1. WHAT WE DID today in plain English.
2. WHAT TO DO at home, numbered.
3. WHAT TO WATCH FOR with the specific phone number.
4. WHAT IS NORMAL recovery vs concerning.
5. THE FOLLOW-UP date.
6. WHEN TO CALL US vs go to emergency.

Use plain English. Specify medications by what the bottle will say, not just generic name. Be specific about timing.

When to use: Before discharge. · Best model: Claude.

Prompt 5

Difficult-Behavior Patient Plan

Aggressive or fear-reactive patients need a structured plan. This prompt drafts it.

Patient (de-identified):

SPECIES, BREED: [BRIEF]
BEHAVIOR PATTERN: [SPECIFIC, what triggers, what they do]
PRIOR VISITS: [BRIEF, what worked, what failed]
MEDICAL CONTEXT relevant to behavior: [BRIEF]
OWNER'S ROLE in the pattern: [BRIEF]

Draft a plan:

1. PRE-VISIT preparation (sedation pre-meds, fear-free protocols, owner instructions).
2. THE WAITING ROOM strategy.
3. THE EXAM ROOM approach.
4. THE HANDLING plan with specific staff roles.
5. THE PROCEDURE-OF-THE-DAY adjustments.
6. THE OWNER COACHING during the visit.
7. THE NEXT-VISIT trajectory.

Safety is non-negotiable. Flag any procedure that should be deferred until behavior protocols are working.

When to use: Before the next scheduled visit. · Best model: Claude.

Prompt 6

Specialist Referral Letter

Specialist referrals need specific information. This prompt drafts the letter.

Referral context:

PATIENT (de-identified): [BRIEF]
REFERRING TO: [SPECIALTY]
REASON for referral: [SPECIFIC]
MY DIAGNOSTIC WORKUP completed: [LIST]
MY DIFFERENTIALS: [LIST]
WHAT I AM ASKING the specialist to evaluate: [SPECIFIC]
TIMING urgency: [BRIEF]

Draft the referral letter:

1. PATIENT SUMMARY factually.
2. WHAT I FOUND and what I am uncertain about.
3. WHAT I HAVE TRIED.
4. THE SPECIFIC QUESTION I am asking the specialist.
5. THE TIMING urgency.
6. WHAT I HAVE TOLD THE OWNER.
7. CONTACT for follow-up.

Do not over-claim diagnostic certainty. The specialist appreciates calibrated uncertainty.

When to use: When making the referral. · Best model: Claude. Calibrated uncertainty matters.

Prompt 7

Online Review Response (Vet)

Vet reviews can be emotional. This prompt drafts the response without escalating.

Review I am responding to:

[PASTE REVIEW]

MY VERSION OF EVENTS (HIPAA-equivalent considerations apply for pet records in many jurisdictions): [BRIEF]
WHAT I CAN OFFER to resolve: [BRIEF]

Draft a response under 80 words:

1. ACKNOWLEDGE the emotion without confirming we treated the pet.
2. INVITE them to call the office directly.
3. CLOSE warmly.

Do not reveal patient details. Do not get into the specifics publicly. Do not match emotion with defensiveness.

When to use: Within 48 hours. · Best model: Claude. Discipline matters.

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 vets this means AI can read your scheduling, your Gmail thread with referring clinics, your Notion practice workflow library. Keep AI out of direct PMS integration without policy review.

For veterinarians, the connectors worth pairing with these prompts:

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.

Two ways to go further