Building the AI factory: the operating-model shifts behind Doctolib's AI products

Every company is racing to put AI into its products. Far fewer are solving the harder problem underneath: making AI products something every team can build reliably, safely, and at scale. In our case that happens in healthcare, where 100M patients and 580k healthcare professionals depend on what we ship.

As Product Manager of the Data & AI Platform at Doctolib, this is my field report, told from the platform's perspective, on two years of transformations that laid the foundations of an "AI factory":

• From artisanal to platform-powered AI products, and the strategy to get there.
• From platform capabilities to accelerators: taking an AI product to production went from several quarters to a few weeks.
• From manual coding to an AI-native SDLC, and what it means for the platform.
• From technical enabler to platform-as-a-product: treating internal feature teams as customers.

I'll share the impact of these transformations with real numbers, along with our learnings, trade-offs and decisions we got wrong. You'll leave with a concrete map of the operating-model decisions behind scaling AI products, the ones that made sense in Doctolib's context.