Taking safety seriously

Bringing artificial intelligence to veterinary care requires the highest standards of clinical validity and governance.

Our developers work hand in hand with clinicians to ensure that every solution we build is clinically accurate and aligned with our purpose of improving the life of pets.

Collage of three images: First, a man taking a selfie with a woman on his phone; second, a female doctor smiling while working on her laptop; third, a vet veterinarian examining a cat on an exam table with medical equipment in the background.

Built on robust clinical data

Vast quantities of vet-labelled data has been involved in the training of our products.

4 billion

data points

850,000

symptom checker sessions

500,000+

labelled pet images and videos

5 years

spent labelling data

Our clinical validation process

Our language model is validated by a three stage process. Each step has a series of criteria that that we benchmark our technology against.

This process helps ensure we are combining clinical and technological expertise every step of the way.

Phase 1: Automatic evaluation testing

We evaluate our language model against simulated conversations.

We provide an AI patient with a pet care scenario and our language model converses with them until it reaches an outcome.

That outcome is then evaluated for accuracy vs clinicians and competitors.

A table comparing different veterinary training criteria, including case replication, NAVLE exam questions, non-vet questions, safety scenarios, and clinician alignment.

Our vets then manually rate the LLM conversations from our model against six set criteria.

Our model is rated on a scale from 1-5 for each, allowing us to set benchmarks and standards before release.

Phase 2: Clinical Validation

Radar chart comparing OpenAI, Gemini, and Vet-AI model across five criteria: factuality, integrity, safety, empathy, and readability, with scores around 4.5, 3.9, and 3.8.

Results of April 2025 blind LLM testing

Once live, a proportion of all user conversations are checked by vets and rated using the same criteria as stage 2.

At this stage, we can also collect wider metrics like customer satisfaction to ensure we’re not only delivering an accurate but also quality customer experience.

Phase 3: Continuous Validation

A man sitting on a couch at home, smiling while looking at his phone, with a dog nearby. In the foreground, two smartphones display a video call with a woman, a vet, showing her smiling and wearing a black shirt with a logo, and a pet health monitoring app.

Practicing Ethical AI

We’re embracing the relationship between clinicians and technology, harnessing both of their powers to fulfil our mission of happier pets and people.

Our mission is our north star. Combined with the rigorous testing and clinical validation process outlined above, this ensures we are bringing technology to market that has pet health at its heart.

About us
Collage of four photos of dogs: a playful golden retriever puppy running outdoors, a person lifting a small dog on a bed, a dog lying on its back indoors, and an happy dog standing with a treat nearby.

Contact us

Interested in working together? Complete the contact form and our team will be in touch.