This marks a breakthrough that will enable more pets to receive the right care faster and at a lower cost. It comes at a time that the profession faces workforce shortages and rising costs, alongside growing scrutiny from the UK’s Competition and Markets Authority.

Independent testing proves performance advantage

In a blind testing conducted in April 2025, UK vets reviewed 48 simulated pet-owner conversations generated by Vet-AI’s model, Gemini 2.0 and ChatGPT 4.0.a. Each was scored for clinical accuracy, triage effectiveness, and safety, empathy, factuality and readability. Vet-AI’s LLM achieved:

• 81% clinical accuracy, outperforming Gemini (69%) and ChatGPT (50%)
• 81% correct triage stops, ahead of Gemini (75%) and ChatGPT (56%)
• Higher ratings for empathy, safety and thoroughness in responses to owners.

“This performance matters because most vet appointments, whether it’s for skin complaints or an upset stomach, don’t always need in-person visits,” said Sam Webster, Chief Veterinary Officer, Vet-AI. “By accurately identifying what needs urgent care and what can be managed remotely, our technology frees up vets to focus on the most complex cases, reduces unnecessary trips for owners and helps insurers control rising payouts.”

Built on real-world veterinary experience

Vet-AI’s LLM employs a multi-agent design, comprising a ‘receptionist’ AI for initial triage, specialist agents for common conditions, and full clinician oversight, thereby blending LLM performance with medical-grade safety.

It is trained on more than 400,000 vet video consultations and over 4 billion data points from licensed UK vets. It now supports 12,000+ real-world interactions every month and powers:

• Emergency detection, which flags serious symptoms to owners before they’ve even started a consult
• Summaries for vets before consultations, saving time and allowing clinicians to focus on diagnosis and treatment rather than admin
• 24/7 advice through the Joii Pet Care app, used by 465,000+ registered owners with a 95% satisfaction rating

Every case handled by Vet-AI’s platform is reviewed weekly by veterinary teams, ensuring its accuracy improves as the model scales. This human-in-the-loop approach delivers continuous, real-world feedback, enabling the AI to evolve responsibly while meeting clinical governance and regulatory requirements.

Paul Hallet, Vet-AI Founder and Exec Chair, added , “This is about resolution, not deflection. Vet-AI’s mission has always been to make high-quality pet care more accessible and affordable for every owner. With vet bills up 60% since 2015 and one in five pets unregistered with a practice, many owners delay treatment or turn to “Dr Google” for advice.

“By helping catch problems early and guiding owners with accurate advice to the right outcome, we can improve care, reduce costs and create a more sustainable model for the whole industry.”

For more startup news, check out the other articles on the website, and subscribe to the magazine for free. Listen to The Cereal Entrepreneur podcast for more interviews with entrepreneurs and big-hitters in the startup ecosystem.