Rafael Laffarga is a seasoned digital leader with over 25 years of experience delivering cross-industry data and analytics solutions globally. Over the last 15 years in the biotech and pharma industry, Rafael has specialized in transforming complex business functions—including HR, Finance, and R&D—by shifting the focus from outputs to tangible, value-driven outcomes. A passionate advocate for data sharing and data governance, he currently spearheads strategic initiatives that bridge the gap between robust data foundations, cutting-edge AI, and frictionless employee experiences. Rafael holds an Engineering degree, an Executive MBA from IE Business School, and a Digital Excellence Diploma from IMD in Lausanne.
Data platforms weren't built for agents that act, decide, and operate on sensitive data. Yet across many organisations, that is already the reality: data products are being consumed not just by people, but by AI systems working across domains like people and finance. At Roche, Rafael Laffarga is rethinking how data products and platforms are designed so they can support this shift. With a growing catalogue of governed data products and federated ownership in place, the focus is moving towards making these products usable, traceable, and safe for both human and agent-driven consumption, without slowing down value creation across the business. This includes tackling new questions around ownership, lifecycle, and cost, as well as understanding what agents are interacting with data, how they behave, and how their actions can be tracked and improved over time.
Data teams are increasingly using AI tools and agents to accelerate development, documentation, and analysis. While productivity gains are real, they introduce new challenges around governance, security, context, and cognitive load. This panel explores how organisations enable data teams to use AI effectively while maintaining quality, control, and sustainable ways of working.
Applying AI safely to improve data team productivity
Managing risk, access, and oversight when using AI tools
Scaling AI usage across teams without increasing fragility