clinical monitoring


Rapid evolution: How agentic AI is redefining the role of the CRA
The way clinical trials are monitored is once again about to change.
This isn’t the first time clinical trial monitoring has evolved. Over the past three decades, the industry has undergone two major transformations in how monitoring is performed.
For much of the 1990s and early 2000s, monitoring relied on frequent on-site visits and extensive source data verification (SDV), with many studies aiming to verify nearly every data point. While rigorous in intent, this approach became increasingly difficult to sustain as trials grew larger and more complex, delivering diminishing returns relative to its cost and operational burden.
The industry responded by adopting risk-based monitoring (RBM) and centralized monitoring, shifting from exhaustive verification to a targeted, data-driven approach focused on the risks that mattered most. This evolution was reinforced by FDA and EMA guidance and ultimately codified in ICH E6(R2) in 2016.
While RBM improved efficiency and data quality, it did not fundamentally change how monitoring work was performed. CRAs still spent much of their time manually reviewing data, reconciling information across systems, documenting findings, and coordinating follow-up activities.
We are now entering a third shift. Unlike the first two, which primarily redistributed how monitoring effort was allocated, this one fundamentally changes the old rules on who and what is monitoring trial performance.


