Veterinarian — tasks, one by one
The unit of analysis is the task, not the job title. Each one below carries its direction, whether the judgement rests on evidence or on platform inference, the reasoning, and what it does not establish.
Every task on this page#
Examining the animal
Still human-led≈ Platform inferenceHandling and examining an animal that cannot say where it hurts, and taking the history from an owner who may have missed things.
The examination is hands-on work on a frightened or uncooperative patient — palpating, listening, watching how it moves — combined with questioning an owner. Wearables and apps can record activity or flag changes at home, but they add information before the visit rather than doing the examination.
This rests on the nature of the work and the state of the technology, not on a record of how examinations are done; it says nothing about how many consultations move to video calls for simple questions.
Reading X-rays, scans and test results
Being augmented≈ Platform inferenceInterpreting radiographs, ultrasound, blood work and slides to reach a diagnosis.
Software that reads animal radiographs and counts cells on slides is sold to veterinary practices, and laboratory analysers already produce most numbers without a person. The vet still decides what the findings mean for this animal; in Singapore, a Veterinary Practice Act passed in April 2026 lists diagnosis, including tests done to diagnose, among the acts only a qualified vet may do once it comes into force.
This rests on what the tools can do, not on a measurement of how many practices use them or how accurate they are on real cases; and the new law is not yet in force.
Treating, vaccinating and operating
Still human-led✓ Evidence-backedGiving treatment and vaccinations, performing surgery and dentistry, and managing anaesthesia and recovery.
This is skilled hand work on living animals of very different sizes and species, which current robots do not do. In Singapore the law also says that, apart from authorised officers, a person must not treat, vaccinate or inoculate any animal or bird except under a licence issued by the Director-General and its conditions.
This is one country's law and the state of the technology; the law establishes who may treat, not that tools cannot assist, and it says nothing about how the work is shared with veterinary nurses.
Prescribing and signing certificates
Still human-led≈ Platform inferenceChoosing and prescribing medicines, and signing health, vaccination and travel certificates that others rely on.
A prescription and a certificate are decisions someone answers for: an airline, an importing country or a pharmacy relies on the signature. Software can suggest doses and fill in forms, but the signing stays with the vet; Singapore's new Veterinary Practice Act lists prescribing and certificates that require professional judgement among the acts reserved to qualified vets once it is in force.
The new law is not yet in force and its detailed rules are still to be made; this judgement does not establish how much of the drafting behind a certificate software already does.
Talking it through with the owner
Still human-led≈ Platform inferenceExplaining what is wrong, the options and their cost, and helping owners decide — including when to end an animal's life.
These conversations carry money, guilt and grief, and the owner has to trust the person recommending treatment or euthanasia. Chat tools can answer general questions before and after a visit; the decision about this animal is made with the vet.
This rests on the nature of the conversation, not on a record of how owners use online advice or how often they now arrive with a diagnosis from a chatbot.
Notes, reminders and bills
Automating≈ Platform inferenceWriting up consultations, sending vaccination and check-up reminders, and preparing estimates and invoices.
Practice software already sends reminders and produces invoices without anyone typing them, and tools that listen to a consultation and draft the clinical note are sold to veterinary practices. The vet reads and signs the record; the writing is increasingly done by the software.
This rests on what the software offers, not on a measurement of how many practices use note-drafting tools; errors in a drafted note are still the vet's to catch.