In this talk, Franck Coste (Snowflake) and Simoh-Mohamed Labdoui, Group Head of Data & Analytics at Saint-Gobain, present the group's data and AI strategy — a global leader in sustainable construction with 160,000 employees across 80 countries.
The core conviction: AI can't be plugged directly into unprepared data. Simo walks through the necessary foundations — orchestration, integration, data quality, and access control — before covering three further pillars: easier data access for business teams, business contextualization (essential to prevent AI misinterpretation), and self-service through conversational interfaces and augmented analytics.
The talk includes several concrete examples — a conversational tool for finance data, a group-wide data marketplace to facilitate global data access — along with key architectural decisions: placing the semantic model as close to the data as possible to prevent AI hallucinations, role-based access control (RBAC), and data governance designed around business processes rather than rigid, top-down rules.
