Retrievant is an AI retrieval company. We build agents that go and get a patient's complete medical history, from every hospital, clinic, insurer and filing cabinet it lives in, and hand back one verified record. This page is about who we are and what, exactly, we do.
A person's medical history is scattered. The average adult has seen more than seven providers in the last decade, and each one keeps its own chart. Some are on modern systems with APIs. Most are not: small practices, imaging centres, retired doctors, insurers, and hospitals in another country. Getting the whole record has meant weeks of phone calls, faxes and waiting.
Retrievant replaces that work with software. A patient authorises once. Our planning model works out where they've been treated and the fastest way to reach each place. Then our agents go: direct API queries where they exist, browser agents that log into portals, voice agents that phone the front desk, and a fax intake pipeline for everything on paper. A clinical reader structures what comes back, a reconciler merges and de-duplicates it, and every field is delivered with the source it came from.
We sell this to organisations that need complete records to do their job: injury and disability law firms, life and health underwriters, clinical research teams, clinics onboarding new patients, digital health products, and cross-border care providers. They pay per completed retrieval. Patients never pay, and can see and revoke every authorisation from their own dashboard.
Co-founder & CEO
Led product for the AI features inside France's largest appointment and records platform, used by most French GPs. Previously scaled a digital health insurer from launch to a million members. Trained as a clinical data engineer.
LinkedIn →Co-founder & CTO
Built hospital interoperability infrastructure connecting 700+ facilities, then led ML engineering at a biotech training models on real-world oncology data. Has written more HL7 parsers than he would like to admit.
LinkedIn →Élodie and Julien leave their roles and start prototyping a voice agent that can request records from a French GP practice. It works on the third clinic.
A Boston personal-injury firm and a Paris clinic go live. Browser agents and fax intake are added because the phone alone isn't enough.
The US is where retrieval is most broken and most expensive. We open in Boston, hire the first retrieval specialists, and cross 30,000 reachable care sites.
Pre-seed backing from Nolqa to extend agents across the remaining US states and into Germany and Spain, and to publish our work on agent safety in health contexts.
Not the model, not the dashboard. If a field is wrong or missing, nothing else matters. We measure ourselves on completeness and provenance first.
Every new capability starts with a human in the loop and only graduates when the numbers say it's ready. We'd rather be slow than sorry with health data.
Our agents disclose they're automated on every call. Patients see everything done in their name. Trust is the only asset a health-data company really has.
Want to join? We're hiring across all four teams.
Bring a real case with a messy history. We'll run the agents live on a 30-minute call.
Book a demo