Another "Charlie" at the Hospital: the GHT Rouen Cœur de Seine Rolls Out Its Everyday AI

The name may raise a smile here at Emerit Science: the GHT Rouen Cœur de Seine has also named its artificial intelligence "Charlie". Behind this coincidence lies a concrete hospital digital-transformation project, designed to equip caregivers day to day rather than to create a showcase effect.

The initiative is worth a closer look, because it illustrates questions we know well at Emerit Science: how do you deploy a useful, secure and sovereign AI in an environment where data confidentiality is non-negotiable? Here's our take.

Rather than building its own chatbot, the hospital group adopted Orange Business's Live Intelligence platform (via the Resah purchasing hub), customized into an aggregator of everyday AI tools that will eventually be open to 18,000 people across its nine facilities — from administrative staff to surgeons, including buyers and technicians.

"Our goal is first to solve 70% of the use cases found in a hospital: drafting an email, correcting a text, working through documents… We'll address the rest with specific projects," explains Thierry Samper, AI project lead at the IT department, who favors real "groundwork" over the "wow" effect.

• Write and edit: draft an email, correct a text, rephrase a note.
• Summarize: condense and make use of documents.
• A pragmatic approach: prioritizing real "groundwork" and genuine needs over the "wow" effect.
• A controlled framework: giving staff a calmer, more controlled AI environment.

The choice of name is no accident: it refers to Charles Nicolle, the French physician who won the Nobel Prize in Medicine in 1928 for his work on typhus, and whose name is associated with the main site of the Rouen University Hospital. A way of tying technological innovation to the institution's scientific heritage.

The coincidence with Emerit Science's Charlie is purely fortuitous — yet it says something about our times: everywhere, healthcare players are seeking to give a familiar, reassuring face to AI tools set to become part of teams' daily lives.

A notable point for a public project: security was treated as a priority, with several structuring safeguards.

• Controlled access: single sign-on (SSO) login on the institution's network.
• Closed circuit by default: requests stay, by default, within a controlled perimeter (users can open them to the Internet when needed).
• Nearby hosting: data hosted 30 kilometers from Rouen.
• French trusted third party: models trained on Orange's servers, chosen as the trusted operator.

So many choices that echo the conviction we defend: AI in healthcare is only valuable if it respects the confidentiality and sovereignty of the data it handles.

The rollout comes with genuine upskilling work: prompt engineering workshops help staff frame effective requests, and an ambassadors system spreads usage in concentric circles, through peer-to-peer training.

The timeline is gradual: the rollout began on May 10, with access for all staff targeted for early autumn. Usage stays responsible — a capped monthly credit, with the energy and carbon cost shown for each request — and a library of use cases grows as the hospital's needs evolve.

The Rouen project confirms an underlying trend: healthcare institutions want AI that is useful day to day, governed and sovereign. That is exactly the ground we are advancing on with Charlie, Emerit Science's scientific AI agent.

Where Charlie GHT Rouen equips teams' administrative comfort, Emerit Science's Charlie focuses on scientific and biomedical research — two complementary facets of the same requirement: putting AI at the service of professionals, without ever sacrificing control over data.

Whether it's a hospital group in Normandy or a research platform, the message is the same: the most promising AI in healthcare is the one that stays controlled, compliant and close to teams. The coincidence of names is just a wink; the demand for sovereignty is very real.

At Emerit Science, this conviction guides every design decision behind Charlie — so that researchers and institutions can benefit from AI without ever giving up control of their data.