How UK practices use AI in veterinary medicine

Most of the AI a UK practice can use today works around the consultation rather than inside the clinical decision. It writes up what was said in the room and prepares the paperwork that follows a visit. For a veterinary surgeon or practice manager looking at AI veterinary software for the first time, the useful questions concern where it already helps in daily work, and the checks a practice should make before it goes near a client's records.

Where AI is used in veterinary practice today

The uses that have reached ordinary clinics are mostly about records and information. Here is where AI turns up in a working day.

Clinical notes written from the consultation, with an AI scribe listening to the conversation between the vet and the client.
Dictation, where the vet speaks the notes and the software transcribes and structures them.
Knowledge search across evidence-based sources and research articles chosen with vets.
Discharge instructions for the owner, generated from the clinical notes.
Billing suggestions based on what the notes record, so that fewer charges are missed.
Suggested diagnosis codes drawn from the vet's own note.

Every one of these is a support or administrative task. The software prepares a draft or offers a suggestion, and a veterinary surgeon decides what goes into the record and what happens to the patient. Claims that AI diagnoses animals on its own sit outside what these tools do, and a practice is right to treat such claims with caution.

How an AI scribe works in a consultation

An AI scribe takes on the notes that would otherwise be written up between consultations or after the last one. Vetz describes the process in five steps.

When the consultation begins, the vet starts Vetz and it begins listening, so attention can stay on the animal and the owner.
Vetz interprets the conversation as it happens and works out who said what and which information matters.
Using the clinic's own templates, it has a structured draft clinical note ready within a few seconds of the consultation ending, written in medical terminology.
The vet reads it through and changes anything that is wrong, either by hand or by asking Vetz to rephrase or clarify a passage, and then approves it.
Where the integration is active, the approved note goes into the practice management system with one click. Vetz integrates with Provet Cloud, Vettigo and MHK FileMaker.

Vetz handles conversations with more than one speaker, so the owner and the vet can both talk naturally. It works in more than 10 languages and also covers video and telephone consultations. The page on AI note-taking for vets goes into the scribe in more detail.

Decision support stays in the vet's hands

The second group of tools helps with clinical thinking rather than writing. Vetz has a knowledge search, an AI agent that searches evidence-based sources selected together with vets as well as the practice's clinical record data. Vetz describes it as a guide for the harder cases.

Alongside it, the Vetz AI assistant gives insights based on the patient's history and findings while the vet works. It can point out parts of the note that are missing, such as the history or the diagnosis. It can also suggest differential diagnoses to consider and remind the vet when a revisit or follow-up is due. Each suggestion is a prompt to think again, and the veterinary surgeon decides whether to use it. Vetz describes itself as a support, with the vet always in control of the note.

What the RCVS says about AI

For vets in the UK, the professional body to read on this is the Royal College of Veterinary Surgeons (RCVS).

The RCVS has published advice for the profession on using AI in practice. It notes that the Code of Professional Conduct and its supporting guidance do not refer to AI specifically, and that their existing principles can be applied to AI tools. According to the RCVS, vets and veterinary nurses remain professionally responsible for the decisions made when treating their patients, and it would be inappropriate to wholly delegate clinical decision making to an AI tool, with a human always in the loop taking responsibility for decisions. The same advice describes being open and honest with clients about how AI is used to process data in the practice as the simplest and safest way to avoid compliance issues.

The RCVS supporting guidance on clinical and client records says that veterinary surgeons and veterinary nurses must keep clear, accurate and adequately detailed records, and that entries should be objective and factual and made at the time or as soon as reasonably practicable afterwards. For scribing tools, the RCVS AI advice says this means AI-generated records should be verified manually, with any necessary edits made at the time. When an error is corrected later, the guidance says the amendment should not remove or obscure the original entry, which keeps the audit trail intact.

A practice that brings in AI should read the RCVS guidance as it stands on the day, and make sure every vet using the tool knows what it says.

Data protection for UK practices

A recorded consultation and the notes made from it contain the client's personal data, from their name and contact details to what they said in the room. That makes an AI tool a question for whoever handles data protection in the practice, as well as for the clinical team.

The UK GDPR defines the controller as whoever determines the purposes and means of processing personal data, and a processor as a body that processes personal data on the controller's behalf. Where a vendor processes data on a practice's behalf, Article 28 allows the practice to use only processors that give sufficient guarantees of appropriate technical and organisational measures, and requires a binding contract to govern the processing. Among other things, that contract has to cover processing only on the practice's documented instructions, confidentiality for the people handling the data, security measures, the conditions for engaging sub-processors, and returning or deleting the data at the practice's choice when the service ends.

These are the questions to put to any AI vendor before signing.

Where are recordings, transcripts and notes stored, and in which country?
Which subprocessors handle the practice's data, and for what purpose?
How long are recordings and transcripts kept, and can the practice have them deleted sooner?
Is there a data processing agreement, and can the practice read it before it commits?
Is client data used to train the vendor's models?

Vetz describes its integration with practice management systems as encrypted and GDPR-compliant, and sets out how it handles security on its security page.

Telling clients that AI is in the room

An owner may well notice a phone or laptop recording on the consulting room table, and it is better that they hear about it from the vet first. Vetz recommends that the vet tells the owner that AI is used to support the clinical notes, since being open about it builds trust, and Vetz can supply ready-made wording for the practice to use.

Under Article 13 of the UK GDPR, a practice that collects personal data directly from a client has to tell them, at the time it is collected, who the controller is and how to contact it, the purposes and legal basis of the processing, who receives the data, whether it will be transferred outside the UK, how long it will be kept and what rights the client has over it. If the processing involves any automated decision-making, the client has to be told that as well.

The practice can also add the same information wherever it explains to clients how their data is used.

Risks to watch for

AI in veterinary medicine brings its own kinds of mistake, and each one has a practical check.

The scribe can mishear a drug name or a dose, so the vet reads every draft against what was said before approving it.
A suggested differential or follow-up is only a prompt, so the vet treats it as one and records their own reasoning in the note.
A fast draft makes it tempting to sign at once, and a practice can make reading the note a stated step in its routine.
Suggested diagnosis codes and billing items need the same review as the note they came from, since an error there reaches the client's invoice.
If notes still have to be copied by hand into the practice management system, the time saved on writing goes into pasting, so the integration is worth testing first.

Many of these risks are ordinary record-keeping problems in a new form. The guide to common mistakes in vet record keeping covers the habits that keep a clinical record accurate, however the first draft was written.

What to look for when choosing a tool

During a trial on the practice's own consultations, these are the things worth testing.

Integration with the practice management system already in use, so finished notes arrive in the right place without copying.
Support for the clinic's own templates, so the notes match how the practice already records a consultation.
Handling of several speakers, and of video and telephone consultations if the practice offers them.
The languages the team and its clients actually speak.
How easily a vet can correct a note, by hand or with the AI's help.
A free trial long enough to cover a normal week of consultations.

Vetz runs in the browser with nothing to install, and there are apps for iPhone and Android. It offers a 14-day free trial with no card details needed, and the full list of what it does is on the features page. Practices thinking about how AI fits the wider running of the clinic can read about the AI veterinary management system. Those looking at AI because of the hours spent on notes after evening surgery may find the page on reducing burnout in vet clinics with AI useful.

More about Vetz

If you would like to see how an AI scribe and decision support would work in your own consultations, book a demo of Vetz and we will show it with your practice's templates and practice management system in mind.