The challenge
Finding the right psychologist is hardest in exactly the moment a person has the least energy for it: directories list hundreds of names with no real way to tell who actually fits, "fit" is about far more than a specialty, and the whole search happens while someone is already struggling.
FindPsych set out to fix the matching problem — but in mental health, doing that responsibly is the hard part. Anything touching clinical territory is bound by AHPRA's advertising and conduct rules, must never resemble diagnosis, has to handle people in crisis safely, and carries real penalties for getting it wrong. The product had to feel warm and human while being rigorously compliant underneath.
What we built
We designed and built FindPsych end to end — an AI-native matching product that stays firmly on the right side of a heavily regulated line.
• A conversational AI intake that asks a few plain-language questions about what someone's looking for — clinical focus, the kind of person they'd feel comfortable with, and practical details like location, budget and availability — eliciting preferences, never a diagnosis.
• An explainable matching engine that returns a short list of AHPRA-registered psychologists who fit, each with the reasons behind the suggestion, so the person chooses who to connect with.
• Safety engineered in, not bolted on: a crisis check runs on every message before the AI is ever called — any self-harm or crisis language is caught and routed straight to Lifeline and emergency support — alongside hard guardrails against diagnosis and an 18+ gate.
• Compliance by design: AHPRA-aligned throughout (no testimonials or ratings — trust is built on verified credentials and transparent matching), with an Australian data-residency architecture for sensitive mental-health information.
The outcome
FindPsych launched as a working product at findpsych.com.au — proof that an AI assistant can guide people toward the right care thoughtfully, transparently, and within the rules of a regulated profession. It turns an overwhelming directory search into a short, explainable shortlist, and shows what responsible, safety-first AI looks like in healthcare. Victoria-first, with a model built to extend nationally.