Eighteen months ago the two of us sat in a borrowed office in Paris listening to a recording of our first voice agent asking a GP's receptionist for a patient's file. It got the date of birth wrong, apologised, tried again, and got the records. We have been chasing that feeling since. Today we can say that Nolqa is backing Retrievant with a $555,000 pre-seed investment, and we want to explain what that changes.
What we do, in one paragraph
A person's medical history is scattered across every clinic, hospital, insurer and filing cabinet they have ever touched. Getting all of it has meant weeks of phone calls and faxes. Retrievant replaces that with agents: a patient authorises once, our planner works out where they have been treated, and voice agents, API connectors and a document reader go and get the records. What comes back is one verified file with the source on every line, usually in under forty minutes.
Where the money goes
We are a small team and we intend to stay one for a while. The investment buys three specific things.
- More retrieval specialists. The bilingual people who take over when an agent leaves its script are the reason our success rate keeps climbing. We are doubling that team in Paris.
- Coverage. Voice and document agents are live in 42 US states and six EU countries. The money extends them to the remaining states and to Germany and Spain, where we already have customers waiting.
- Publishing our safety work. We have learned a lot about consent-scoped agents, disclosure on automated calls and when to escalate to a human. We think the industry should have it, and we are hiring one person to write it up properly.
What it does not change
We are not going to spend it on a bigger model. The hard part of retrieval was never the model; it is the ten thousand ways a records request can fail, and building agents that handle each one. We are not going to start training on patient data. We are not going to stop disclosing that our agents are automated, on every call, in the first sentence. And we are not going to bill anyone for a retrieval that comes back empty.
Why Nolqa, and why now
We spoke to a number of investors. Most wanted to talk about the model. Nolqa wanted to listen to the calls. They asked about hold music, about fax machines in Vermont basements, about what a receptionist says when she realises she is talking to software. Those are the right questions, and a partner who asks them is one we can work with for a long time.
We took the money now because customers are asking for regions we cannot yet serve, and because the specialists we want to hire exist today and may not tomorrow. It is enough to do the three things above well. It is not enough to do anything badly, which is roughly how we like it.
Thank you
To the receptionists who have been patient with our agents, the customers who trusted us with their hardest cases, and the patients who let us fetch their records: this is for you. If you would like to see what we do on one of your own records, book thirty minutes.
Élodie Moreau and Julien Lefèvre
Co-founders, Retrievant · Paris
About Retrievant
Retrievant is an AI retrieval company headquartered in Paris. Its agents retrieve a patient's complete medical history on the patient's authorisation and deliver one verified, source-linked record to legal, insurance, research, provider and digital-health organisations. Retrievant is SOC 2 Type II, HIPAA, GDPR and HDS certified.