Blogs · July 2026

Is it safe to let AI answer calls from pet owners?

Only if it cannot judge how urgent a case is. Here is the line that keeps an AI agent safe in a veterinary practice, and the questions to put to any supplier.

The short answer: yes, but only if the agent is built so that it cannot give clinical advice or decide how urgent an animal's condition is. A safe agent does the admin: it books, it reschedules, it repeats your published fees, it gives your opening hours and policies, it handles repeat bookings. The moment a caller describes symptoms, or anything that sounds urgent, it stops, says so, and gets them to a person on your team. If a supplier cannot state that plainly, do not put them on your phone line.

The risk is the scope, not the technology

The scenario every vet imagines is the same: an owner rings about a collapsed dog and a machine tries to be helpful. That is a real risk, and the answer is not to make the AI cleverer. It is to make its scope narrower and enforce it.

So the question for any supplier is not "how good is your AI?" It is: what is it forbidden from doing, and what does it do when it does not know?

What HealthCentre will not do

  • It does not judge urgency. Not symptoms, not severity, not "does this need to be seen today".
  • It does not give clinical advice, on treatment, medication, dosage or aftercare.
  • It does not guess. It answers only from your own knowledge base: your services, your consult fees, your opening hours, your policies. If the answer is not there, it does not invent one.
  • It does not delay an urgent caller. Urgency routes to a human immediately.

What it does instead

  • Answers every call, email, chat and message in seconds, in your practice's voice, day and night.
  • Books, moves and cancels consults in your real diary.
  • Answers the admin questions that make up most of your phone traffic: what a consult costs, are you taking new registrations, what are your hours, where do I park, when is the booster due, can I move Thursday.
  • Handles repeat bookings and reminders, and clears the inbox overnight.
  • Hands over to a person, and tells the caller that is what it is doing.

Handing over to a person is a designed route, not a failure

Treat the handover as the feature, not the breakdown. You decide what gets passed to a person, to whom, and how quickly. An owner with a frightened animal should reach a person fast, with the details already captured so they do not have to repeat themselves while panicking.

This is also where your obligations sit. Under the RCVS Code of Professional Conduct, vets in practice must provide or arrange 24-hour emergency first aid and pain relief. An AI agent is part of how an owner reaches that route quickly and calmly. It is never a substitute for it, and any supplier who implies otherwise is selling you a liability.

Why owners ring in the first place

Worth understanding, because it tells you what the agent is actually for. Most calls are information calls. Fewer than 40% of practices publish prices on their website, the CMA found, and fewer than half of owners get pricing information before non-routine treatment. Owners cannot find out, so they ring. An agent that answers those questions accurately at 9pm is not doing anything clinical at all. It is doing the job your website could not.

Who is accountable

You are, and that does not change. The knowledge base is yours and you approve it. You set what it answers and what it passes to a person. Every conversation is logged and readable, so if the agent drifts, you will see it and can correct it. An AI agent you cannot audit is not one to put in front of owners.

The questions to ask any supplier

  1. Can it decide how urgent a case is, and if you say it cannot, how is that enforced rather than merely intended?
  2. What happens when it does not know the answer?
  3. How does an owner with an emergency reach a human, and how fast?
  4. Where do its answers come from, and who signs them off?
  5. Can I read a full transcript of every conversation?
  6. Does it book into my system, or just take a message?
  7. Where is client data stored, and what happens if we leave?

Vague answers are the answer.

Frequently asked questions

Can an AI agent decide how urgent a sick pet is?

No, and it must not. Judging urgency is a clinical decision. A safe agent recognises it needs a person and puts the caller through immediately.

What happens if someone rings with an emergency?

It identifies the urgency, tells the caller it is putting them through to a person, and routes them according to the rules you set. It does not attempt to advise.

Does it pretend to be a person?

No. It introduces itself for what it is.

Can it answer questions about prices?

Yes, from your published fees. That is admin, not clinical, and it is what most callers want.

Sources

See it answer for your own practice

Enter your website address and hear HealthCentre answer live, built around your real services, consult fees and hours.

Get started