2026 DIA Global Annual Meeting: Why Protocol Digitization Has Been So Hard to Automate—Until Now

Full interview summary

In a recent video interview with Applied Clinical Trials at the 2026 DIA Global Annual Meeting, Angie Maurer, vice president of AI-enabled clinical development at Medable, discussed why protocol digitization has remained one of the most persistent bottlenecks in clinical trial startups and what has finally changed to make meaningful automation possible. She identified three structural barriers that historically made the problem intractable: the absence of a standard data model, the inherent ambiguity of clinical language, and the industry's zero error tolerance requirement that kept sponsors locked into manual processes. The convergence of LLMs capable of interpreting nuanced clinical language, the USDM standard developed by TransCelerate and CDISC, and the availability of validated operating environments has, she argued, created the conditions to finally solve the problem.

Maurer went on to explain how transforming static protocol documents into structured data fundamentally changes how protocol amendments are managed. In the current model, a single amendment triggers a cascade of manual updates across EDC, IRT, site training, and other systems—a process rife with handoff errors and delays. When a protocol is digitized, each element becomes a data object with defined downstream relationships, enabling automated propagation of changes across systems simultaneously. Layering AI agents on top of this structure, she argued, transforms the protocol from a human-interpreted document into a reliable instruction set that systems can execute against—creating the connective tissue between upstream protocol definition and downstream clinical data infrastructure.

She closed by connecting protocol digitization directly to the FDA's push for real-time data review and continuous regulatory oversight, framing it as the critical upstream enabler that makes everything downstream possible. When protocols are stored as structured, CDISC-compliant data from the outset, every amendment, endpoint, and visit schedule becomes a versioned data point that can be validated, transmitted, and reviewed in real time—turning what was once a startup efficiency problem into a foundational capability for the continuous regulatory environment now taking shape.