This blog is a summary of a podcast Medable sponsored that occurred on May 21, 2025 and was hosted by Emerj.

You can find it here, https://podcast.emerj.com/how-ai-is-transforming-clinical-trials-and-data-access-with-mathew-paruthickal-of-sanofi

 

Artificial intelligence continues to move from experimentation to execution across the life sciences industry. On a recent episode of the AI in Business podcast, Matthew Peruhakal, Global Head of Data Architecture, Utilization, and AI Engineering at Sanofi, offered a deep look at how the pharmaceutical giant is integrating AI to transform clinical trials. From intelligent data workflows to proactive risk detection and regulatory alignment, Peruhakal described an organization reshaping its research and development operations around a new digital core. 

His message was clear: AI cannot remain a side project. To make a meaningful impact, it must be embedded as a strategic capability that connects people, systems, and data across the enterprise.

Clinical data is evolving from digitization to intelligence

A year ago, much of the industry’s digital transformation conversation focused on digitization, such as replacing paper forms with eConsent, implementing electronic data systems, and improving record management. According to Peruhakal, that phase has largely passed. The new focus is on intelligent data integration and connecting structured and unstructured data to drive real-time decisions. At Sanofi, AI is being used to optimize clinical protocol design, predict site risks, and detect early safety signals while maintaining regulatory precision and patient safety.

This shift from basic digitization to intelligent orchestration marks a turning point in how data is treated across the pharmaceutical landscape. It’s not enough to collect information; organizations must make it actionable. “It’s less about the tools and more about how you connect them,” Peruhakal explained. The orchestration of systems defines success,  rather than the isolated use of individual technologies.

Building AI as a strategic capability, not a side project

One of the central points in Peruhakal’s message was that AI must be seen as a strategic capability, not a “shiny side project.” Many organizations fall into the trap of running isolated AI experiments that never scale or connect back to business priorities. For life sciences companies, that approach is particularly risky given the complexity of compliance, data governance, and patient privacy.

At Sanofi, AI is being integrated across data architecture, document management, and content generation. The company’s approach centers on interoperability and creating systems that allow structured clinical data and unstructured documents to work together. Clinical trials generate massive volumes of both: databases full of patient records and a parallel stream of reports, PDFs, and safety documents. AI tools like generative models are being applied to extract context and insights buried in those unstructured files. The result is a more connected and responsive clinical information ecosystem.

Trust plays a defining role in this transformation. In life sciences, every action must be traceable and compliant. “Trust is equally as important as the technology itself,” said Peruhakal. Building governance and auditability into AI systems from the start ensures that innovation does not compromise compliance or patient protection.

The three pillars of Sanofi’s data architecture

Sanofi’s AI transformation strategy rests on three core pillars: modern data architecture, document intelligence, and interoperability. The first pillar involves creating a unified data foundation (often referred to as a “data lakehouse” or data fabric) that combines clinical, operational, and research data. This structure allows teams to query data in natural language rather than relying on specialized programming skills. The second pillar, document intelligence, extracts information from diverse document types, from PowerPoint presentations and Excel files to safety reports and PDFs. By applying natural language processing and translation, Sanofi can analyze global documentation in multiple languages, ensuring faster insights across more than 90 countries.

The third pillar connects the first two, linking structured and unstructured data to enable automated content generation. Sanofi is already using this integrated data architecture to generate clinical study reports, safety summaries, and even regulatory submissions. By linking document intelligence with structured data, AI systems can draft compliant materials automatically, saving time and ensuring consistency.

Designing proactive and intelligent workflows

While much of the industry is still focused on reactive workflows (responding to issues after they arise) Sanofi’s goal is to build proactive systems. “We are moving into a world where everything becomes proactive,” Peruhakal explained. The company is developing systems that can flag risks, generate narratives, and prepare submissions before issues occur. Drawing on historical clinical trial data and real-world outcomes, these tools can identify anomalies early and guide teams through preemptive actions.

The vision is for protocols that can effectively “write themselves,” adapting to prior outcomes and patient populations. Instead of waiting for audits or safety events, the system sends early warnings and recommendations. This proactive approach not only improves operational efficiency but also enhances patient safety, which is the ultimate priority in clinical research. Importantly, Peruhakal was quick to clarify that this shift isn’t about replacing people. Rather, it’s about giving scientific, compliance, and legal teams intelligent tools that surface insights faster and reduce risk so humans can focus on the most important decisions.

Making AI accessible, scalable, and outcome-driven

For AI to succeed at scale, accessibility and scalability must be built into its design. Peruhakal emphasized that AI should not live only within the data science team. Instead, it must be embedded into the everyday tools that employees already use, from clinical systems to marketing workflows. By connecting AI to existing interfaces and ensuring the results are explainable, Sanofi is fostering trust and adoption across business units.

Scalability, he noted, comes from platform thinking rather than one-off projects. This means creating reusable components for model governance, version control, and compliance from day one. Every model and data connection is built with traceability and auditability in mind. The third dimension of success, and arguably the most important, is alignment with business outcomes. AI initiatives must tie directly to measurable impact such as fewer protocol amendments, faster safety reviews, or more efficient audit preparation.

One practical example Peruhakal shared was from Sanofi’s marketing team. Even outside of clinical operations, AI is reducing lead times for regulated advertising. Previously, creating a single ad could take months due to medical and legal reviews. With AI-powered systems that automatically verify regulatory language and claims, those processes can be accelerated without compromising compliance. This underscores AI’s potential not only in research but across the full pharmaceutical value chain.

The cultural dimension of AI adoption

Beyond technology, much of what Peruhakal described comes down to culture. Scaling AI successfully requires shifting organizational mindsets away from experimentation and toward sustained value creation. Teams must understand that AI is not a novelty but an enabler of better, faster, and safer decisions. Aligning every project to a clear business or patient outcome helps avoid the pitfalls of innovation for its own sake.

Responsible AI also plays a key role. The systems being developed at Sanofi are grounded in data transparency and human oversight. As Peruhakal explained, the goal is not to automate away expertise but to amplify it, allowing human judgment to remain at the center of clinical and regulatory decision-making.

Conclusion

Sanofi’s work offers a powerful example of how large pharmaceutical organizations can move from digital transformation to true AI-driven innovation. By integrating structured and unstructured data, embedding compliance from the start, and focusing on proactive intelligence, the company is redefining how clinical trials are designed, monitored, and managed.

For life sciences leaders, the lesson is clear. AI is no longer a side initiative. Instead, it’s a strategic capability that must be woven into the fabric of enterprise operations. Building trusted, interoperable systems today will determine which organizations lead tomorrow’s era of intelligent drug development.

‍