“If you use it well, it’s a teammate”

Artificial intelligence (AI) is rapidly making its way into our lives. What can AI mean for veterinary practice? And should we be happy about that?

“Rarely has a new technology been embraced so quickly as AI.” Yannick Bloem is sitting behind his laptop in Silicon Valley. “ChatGPT went from zero to one hundred million users in two and a half months. It took Facebook four and a half years to reach that number nearly twenty years ago.”

It’s seven in the morning and his first online meeting of the day. Bloem is the CEO of CoVet, a US–Canadian company developing AI tools for veterinarians. He previously worked at Apple, among others. “When the AI revolution began, I asked myself: in which field can this truly add value? My sister is a doctor. I often heard her complain about the administrative burden that left her with less time for patients than she wanted. One of my co-founders at CoVet had family in the veterinary world. He said: ‘the problem is at least as big there.’ I thought: let’s do something about this.”

They developed, among other things, an AI assistant that automatically generates reports in the patient management system after a consultation — as well as a short lay version for the pet owner. The system is now available in English, French, Spanish, and Dutch. The veterinarian wears a small microphone and must speak out loud about what they are doing and observing. “Otherwise, the assistant can’t register it,” says Bloem. “That turns out to be useful for pet owners too, who otherwise miss half of what happens during the consultation.”

The AI assistant records relevant data and leaves out the rest. “That was the big challenge,” says Bloem. “What’s important needs to be documented properly, in a way that works for the vet — including correct medical terminology and context. That’s why I find it important that almost everyone who works with us has a background in the veterinary field. At the same time, you don’t want anything recorded about the owner’s holiday or mistakes when a nurse briefly walks in with a question about another patient.” According to Bloem, the system is now highly accurate. “Of course, the vet must always perform a check. In two percent of the reports, something is corrected or added.”

Thousands of clinics now use CoVet’s services. “Especially in the US, adoption is rapid,” says Bloem. “Pet owners don’t seem to mind, and veterinarians save a lot of time. They can focus on what matters: examining the animal and communicating with the owner. Moreover, documentation is very important in the US due to the litigious culture. But in the Netherlands and Europe, we’re also seeing strong growth. The rest of the world is following: Australia, New Zealand — we even have our first clients in Dubai!”
 

See AI as a tool that takes over routine work and frees up time for personal attention and communication.

Waiting to See

Some veterinarians are still cautious. That doesn’t surprise Hans Kooistra of the Faculty of Veterinary Medicine. Laughing, he says: “We’re a stubborn profession. I’m quite conservative myself.” Still, the professor of Internal Medicine — a very respected figure among vets — is convinced that AI will take off. “Reducing administration is one thing — and an important one — but other applications have great potential too.” He points to AI-assisted analysis of X-rays and other scans, which has already made major strides in human medicine. Or virtual assistants that not only listen during consultations and generate reports but also provide real-time advice based on the latest research. “If we at the faculty can in turn use the data from consultations, we can make some really exciting progress.” Examples include studying genetic causes of disease, monitoring infectious diseases, or improving antibiotic stewardship. “There’s enormous potential in the data that will become available.”

The faculty is already using AI in many areas to drive innovation — especially in research. Kooistra notes an open and curious attitude. “That’s a good thing. Of course, we must remain critical, but we’re in talks with developers to see how we can build connections — for example, with our diagnostic system Petscan and in our teaching. We want to educate a new generation of vets who understand both the opportunities and risks of AI and know how to use it responsibly.”

Lessons from Other Sectors

The faculty can learn from other fields, notes Carien Duisterwinkel of the AI Labs at Utrecht University and UMC Utrecht, a collaboration platform for researchers and practitioners. “An example from human medicine is a project where AI is used to keep formularies for medicines continuously up to date based on the latest scientific insights. But AI can also help with privacy and data storage issues through what’s called ‘federated learning.’”

“It’s a bit technical,” she explains, “but it means that you don’t have to share your data — the model comes to the data. This way, organizations keep their confidential patient information within their own walls while still benefiting from collective knowledge and model improvement.”

Professionals don’t always follow AI’s advice, Duisterwinkel adds. She refers to studies involving police and regulatory agencies on human-AI interaction. “AI suggested likely escape routes for criminals or inspection sites for regulators. And what did we find? Officers and inspectors are human — they only follow the AI’s advice when it matches what they already thought. Otherwise, they don’t.” Explaining how the algorithm reached its conclusion didn’t help much either. Successfully implementing such transformative technology is therefore not just technical — it’s psychological. It involves habits and behavior. “You must be very clear about which problems AI can help solve, where we find it valuable to use it, and where we draw the line — what we absolutely want to keep human.”
 

The heart of our profession is the balance of animal, human, and societal interests. That will always require human judgment.

Emotion and Opportunity

That’s a question Mark Huis in ’t Veld grapples with daily. He’s a veterinarian, owns two clinics, and is currently focused on his startup Vetintelligence, based in the historic Anatomy Building in Utrecht. From there, he wants to shape the future of veterinary medicine — with AI. And yes, artificial intelligence stirs strong emotions. A recent international survey of 10,000 people found that half fear developments are moving too fast, with too little oversight. Huis in ’t Veld recently spoke to a vet worried about mass job loss in the sector. He understands the concern but mostly sees opportunity. Digitization, automation, and AI don’t have to be threats; they can be powerful allies. “If you don’t use them as a vet, you’ll eventually be replaced,” he says. “But if you use them well, they’re a teammate.”

He’s currently training that virtual teammate, because the need is clear. “Clients and staff expect different things from us than ten years ago. Clients want 24/7 availability — ‘The supermarket’s open on Sunday, isn’t it?’ they say. They’re more assertive and come in armed with Google research. Meanwhile, staff want a healthy work-life balance and part-time hours. That’s fine, but it means we’re heading toward an even greater shortage of vets and nurses — especially since one in five quits within five years. That needs to change.”

Instead of constantly firefighting — dealing with an overloaded front desk, staff burnout, complaints about accessibility and costs — Huis in ’t Veld wants structural solutions. He speaks passionately about his virtual assistant “Sophie,” showing how it works on his phone. The results are impressive, much like CoVet’s system.
 

Better Service

But this is just the beginning. Huis in ’t Veld wants to relieve the veterinary nurse who answers the phone. “When do people call?” he says, checking a chart. “Mostly in the morning. Between 8:00 and 9:00 a.m., 17% of calls to Dutch clinics go unanswered. Even later in the day, the phone remains a major source of stress for staff and frustration for clients.”

In his Utrecht clinics, AI assistant Sophie now answers common client questions via WhatsApp. “Sophie recognizes emergencies, asks follow-up questions, and refers cases when necessary.” Setting it up properly takes time, he admits. “You’re essentially writing thousands of prompts — if the client asks this, the assistant replies that. It’s very precise work. Sophie had learned what counts as an emergency — including, for example, a bleeding wound that can’t be stopped. But one client messaged about her dog having bloody stool, saying, ‘There’s blood coming from the anus,’ and Sophie classified it as an emergency. We’ve since adjusted it — now she asks follow-up questions.”

Ongoing refinement is key. “Every interaction is logged and can be reviewed,” Huis in ’t Veld says. “Once a component is stable, consistency is guaranteed, and real time savings begin.” Now that Sophie integrates with the scheduling system, she can make and reschedule appointments, answer questions like “Which ointment did I use last year?” or “When is my dog due for vaccination?” — even after hours and in multiple languages.
 

Futuristic Vision

Huis in ’t Veld envisions an AI assistant that does much more: supports triage, manages scheduling, briefs vets before consultations, offers tips based on the latest research, ensures prescriptions and invoices are ready, updates inventory, and follows up with owners after appointments. “See it as an extra colleague,” he says. “Not a replacement for nurses, but a tool that takes over routine work and frees up time for personal attention, communication, and professional expertise — the things a computer can never replace.”

It may sound futuristic, but Huis in ’t Veld says it’s close. “We now have nine pilot clinics working with Sophie. This fall, we’ll expand to at least forty more — the early adopters. We want their feedback to see how they personalize the system. Will mistakes happen? Of course! But errors happened in triage before, too — we just couldn’t learn from them as effectively.”

The crucial question: how do you ensure the AI’s advice is correct? “Sophie isn’t based on OpenAI like ChatGPT; it doesn’t scrape the web. We’re very precise about which information is included. If I can link to validated systems like formularies or the Petscan system at the faculty, I know the data is reliable. Then I can diagnose with the latest breed-specific insights — something no vet can keep entirely in their head.”

“My dream,” Huis in ’t Veld says, “is for Sophie to one day go through veterinary school — not to replace vets, but to support them with the newest knowledge and even mentor students during their final year and early career. So that care for both animals and owners gets even better. That’s what drives me.”

He’s not alone. More veterinarians see potential benefits — time savings, better client interaction, more job satisfaction, and more consistent Good Veterinary Practice. Sophie Deleu, chair of the Royal Dutch Veterinary Association (KNMvD), shares that view. “I’m very enthusiastic about what new technology can bring. But we must dare to experiment — and not be naïve.” Quality control is crucial, she says. “AI learns on its own, but its quality depends entirely on the data you feed it. And oversight remains essential. That requires regulation.”

Beneath that lie deeper questions, Deleu adds. “What does this mean for our profession? How will we use the time we save — to deliver better care, with more attention to each patient? What new ethical questions will arise? Will convenience lead to complacency? What does good care even look like in the age of AI? The heart of our profession is the autonomous balance of animal, human, and societal interests. That will always be tailor-made work. You can never outsource that. As AI changes society, we must find that balance together — while the train keeps moving. It’s an exciting time.”
 

This article appeared in: