Building an AI-Powered Symptom Checker for a Pet Wellness Startup
A pet wellness startup had a compelling idea: an app that helps dog and cat owners understand when their pet actually needs a vet and when they just need rest and observation. The challenge was twofold. First, building reliable AI symptom analysis for non-human patients with limited labeled data. Second, designing an interface simple enough for pet owners with zero medical knowledge — while still being medically responsible enough that vets would endorse it.
We partnered with two veterinary consultants to define symptom categories, red-flag conditions, and safe-to-monitor thresholds. We fine-tuned a language model on veterinary case notes and built a multi-modal input system — text descriptions and optional photos. For the photo analysis, we implemented a lightweight CNN that flags visible abnormalities (skin conditions, eye discharge, limping patterns) without requiring full diagnostic precision. Every AI output was designed to recommend professional consultation rather than replace it — building trust with both users and regulators.
We shipped a React Native app with a conversational symptom checker, photo upload for visual assessment, a health history journal, and a vet directory with in-app booking. The AI layer was built on a Django backend with an OpenAI fine-tune plus a custom CV model. We implemented strict output guardrails — the system never diagnoses, it triages. Each recommendation comes with a clear confidence level and escalation path.
The MVP launched with 400 waitlist users. In the first month post-launch, 78% of users said the app helped them make a faster, more confident decision about veterinary care. Three regional vet clinics signed partnership agreements for in-app referrals.
Facing a similar challenge?
Every project we take on starts with understanding your specific constraints and goals. Let's talk about yours.