In April 2026, my team and I took part in the Harvard Health Systems Innovation Lab’s annual worldwide hackathon, where over two days we developed orAIti, a patient-facing LLM-driven triage tool and business model with bidirectional benefits. It equips patients to start their care journeys with confidence while reducing unnecessary ER utilization, and it would be accessible through well-matched NGO partners, providers, and insurers, giving their users a more reliable triage tool than commonly-relied-upon alternatives. Our project won the Coup de Coeur award.
Who gets lost in the system
Patients with low health literacy face a three-part breakdown. They can’t assess what their symptoms mean, they don’t know who to contact or when, and they can’t advocate for urgency when they do. That drives two costly failures, delayed care for serious conditions and emergency departments filled with patients who needed a GP rather than a hospital. France recorded around 21.3 million emergency department visits in 20241, regional health agency estimates suggest 30 to 40% of those patients could have been cared for elsewhere2, and each visit costs the public system about €2273.
How it works
A fine-tuned LLM handles intake, letting the patient describe what’s happening in their own words and mapping it to structured clinical data, with follow-up questions grounded in what clinicians actually ask. A rule-based triage layer then judges how serious it is, either telling the patient to go now and why, or explaining what to watch for and when to escalate. Finally, an LLM drawing on validated sources builds an action plan covering what to do now, what to say to their doctor, and how to advocate for themselves if they’re dismissed.
Where I came in
Everything in this project was a genuine team effort, and the ideas grew out of conversations we all shared. Within that, I contributed to identifying the weaknesses of our competitors, the free general-purpose LLMs many people already turn to with health questions, and the concerns patients have when using them, along with who feels those pains most, parents being one clear example. I helped shape how we positioned orAIti as a patient education tool that also brings major financial and operational benefits to hospitals and insurers, which became the heart of our value proposition, and I helped identify potential partners and investors who might be interested. Caitlin Chesnais-Dixon and I brainstormed the revenue model together, where we decided on the insurance and NGO channels. orAIti would be funded through NGO grants and paid licences for insurers, employers and hospital networks.
Two days, many pivots
My first hackathon was an exciting trial-by-fire in understanding how product development happens in this space. With each round of mentor feedback, our project evolved and pivoted rapidly. Keeping the patient experience at the core, especially for those with lower health literacy, proved challenging while balancing all the considerations that go into generating viable health tools. We succeeded in centering patient needs in our value proposition, which led us to win the Coup de Coeur award.
Built with Antoine Chesnais, Caitlin Chesnais-Dixon, Lorena Gómez Acosta and Mélissa Laine, and with thanks to Sia Togarrati, whose personal experience helped ground our understanding of what patients need.