Training 50 AI Champions to scale AI

In late 2024, Doctolib noticed a paradox: AI already powered its products, but internal usage remained fragmented. Some teams used ChatGPT on their own, others didn't dare get started. Adoption was uneven, with no shared framework.

Rather than limiting itself to rolling out a tool, Doctolib chose a product approach to its AI transformation. A cross-functional task force brings IT and People together to run the program, with joint sponsorship from the CISO/CIO and the Chief People Officer.

A three-phase strategy

The program follows a clear progression:

Phase 1, individual efficiency: access to Dust as the AI orchestration platform, horizontal assistants (intranet, translation, writing), training in generative AI fundamentals

Phase 2, team impact: departmental assistants connected to business data (Sales, Marketing, People, Finance, Tech), hackathons aligned with OKRs

Phase 3, company-wide ROI: structured impact measurement, workflow automation, assistants built into critical processes

The key role of AI Champions

At the heart of the program, 50 AI Champions spread across the organization act as relays between the central task force and teams on the ground. But being an AI Champion isn't improvised: it takes knowing how to identify the right use cases, structure a request, and above all bring along colleagues who are sometimes skeptical.

Hymaia stepped in to equip these champions, running two workshops focused on the Product mindset applied to AI and the techniques for driving adoption.

Concrete results

Six months after the program launched:

32% active users in Q2, up from 22% in Q1

A Sean Ellis Score of 45% (past the product-market fit threshold)

Measurable gains: 2 to 3 hours a week saved on SQL queries for tech teams

A target of 100% active users by year end

The principles that make the difference

Three convictions guide the program:

AI literacy over mastering one tool: training people to understand generative AI, not to use one specific platform. The skills will transfer beyond today's tools.

Critical thinking as non-negotiable: every assistant carries a disclaimer, and copy-pasting without review is forbidden. AI is compared to a brilliant intern who still needs supervision.

Bottom-up innovation: the program structures the top-down effort but celebrates peer-to-peer learning and the assistants teams build on their own.