How are companies distinguishing between “moonshot” automation and validated digital tools that can be deployed today?

Two classes of technology, two different risk categories

BioPharma companies are navigating two fundamentally different classes of technology: audit-ready, validated systems built for defensibility and repeatability, and fast-moving AI and automation tools optimized for speed and exploration. According to Marc, the real question is no longer whether to adopt these technologies, but which problems they are fit to solve—and the answer has less to do with technical capability than with trust and liability.

Where humans remain in the decision loop and the consequences of error are limited, AI can accelerate experimentation and insight generation. But where decisions carry personal legal responsibility—a Qualified Person under EU and UK medicines law, for example—the bar for validation rises dramatically. Treating these two categories as interchangeable is not merely a compliance risk; it is a strategic misjudgment.

AI derives its value from validated systems—it does not replace them

Marc emphasizes that early hopes of AI simply working around poor legacy data have largely faded. The workhorse foundation—validated ELNs, LIMS and regulated data platforms— still matters. AI derives its value from those systems; it does not replace them. As he puts it: “The moonshot sits above the mandate, not instead of it.”

Governance must keep pace with adoption

Tools introduced quickly in R&D can inch closer to regulated decision-making without a corresponding reassessment of validation and accountability. Governance must evolve at the same pace as adoption. Before any tool goes live, Marc argues, one question cuts through the complexity: who is accountable if this is wrong — and what happens to them?

Read the full roundtable on Pharma’s Almanac.

Discover how IDBS supports compliant data foundations for BioPharma: Explore the IDBS GxP Resource Hub.