AGENTIC AI FOR PREDICTING AND PREVENTING DATA DRIFT

This paper presents a human-centered approach to agentic AI as a novel solution for the prediction and prevention of data drift in DCTs using real-world, available datasets from wearable devices and sensors.In this approach, continuous monitoring of evolving data patterns is done to safeguard the integrity of clinical trial results.It automatically minimizes human intervention with the help of corrective mechanisms and cuttingedge machine learning methods, while allowing for quick response to changes that may take place unexpectedly in the distribution of data.We outline the implementation process, describe a comparison between our method and classic data quality management techniques, and outline some of the challenges-including regulatory issues and biases-that need to be overcome.These results indicate that the use of agentic AI significantly improves data reliability, thereby presenting new avenues for more accurate and efficient DCTs.