The Tufts Center for the Study of Drug Development (CSDD) and Medable, a clinical trial platform company, have released an analysis suggesting that Medable’s Clinical Monitoring Agent could deliver a significant return on investment. The study estimates an 82x ROI for phase 3 trials and 64x for phase 2, translating to roughly $21 million in added financial value for phase 3 and $7.5 million for phase 2.
This analysis comes as the use of AI agents in enterprise settings is on the rise. According to McKinsey, 93% of surveyed organizations reported exceeding their AI budgets. Meanwhile, IBM found that 41% of software developers saved one to two hours daily using AI agents.
AI Agents in Clinical Trials: A Growing Trend
In the clinical trial domain, a 2025 Tufts CSDD analysis of 36 AI/ML use cases found an average 18% reduction in cycle time, with the largest gains in patient monitoring and enrollment assessments. The Tufts-Medable collaboration aimed to translate these operational improvements into investment terms.
Medable’s Chief Medical Officer, Dr. Pamela Tenaerts, highlighted the potential for sponsors with large oncology portfolios, projecting an incremental eNPV of up to $226 million for those with 20 active indications and $565 million for those with 50.
Medable’s agents are prebuilt but require customization. “About 20% tweaking is needed,” Tenaerts noted. These agents integrate with various systems, including EDC, eTMF, and CTMS platforms.
Handling Complex Data in Clinical Trials
Phase 3 protocols now involve an average of 5.9 million data points, up from 929,000 in 2012.
Clinical trials face high failure rates. A study in the Journal of Medicinal Chemistry found a 95% cumulative failure rate for small-molecule cancer drugs from phase 1 to phase 3. Even in phase 3, the success rate is only 43.4%.
AI agents can help by spotting issues early. This allows researchers to focus on higher-priority decisions, as Getz pointed out.
While AI adoption in clinical trials is still early, Tenaerts sees potential beyond oncology. However, full autonomy remains limited.
The Role of AI in Clinical Trial Optimization
Medable’s Clinical Monitoring Agent is designed to identify issues such as enrollment delays, protocol deviations, and adverse events earlier than human researchers.
Current Limitations and Future Potential
The success rate of phase 3 oncology drug trials is currently 43.4%, according to a 2026 analysis. By improving data management and issue detection, AI agents can help increase this success rate and accelerate the development of new treatments. As the technology matures, it is likely that AI will become an essential tool for clinical researchers, enabling them to make more informed decisions and ultimately improve patient outcomes.
AI Agents’ Autonomous Capabilities and Challenges
Frontier AI labs have demonstrated models performing complex tasks with minimal oversight. Anthropic’s Mythos conducted autonomous genomics research over a week, while Google DeepMind’s AlphaEvolve optimized algorithms at a scale beyond human capacity. These examples highlight advancements in unattended operations.
However, challenges persist in integrating such efficiency into existing systems. Dr. Pamela Tenaerts noted that accelerating monitoring, such as sending emails to sites, risks overwhelming recipients if downstream processes remain unchanged. She emphasized the need to address bottlenecks across the entire system to avoid shifting inefficiencies.
Clinical trials face optimization hurdles due to randomization limitations. Testing alternative endpoints, eligibility rules, or doses typically requires new trials, demanding additional time and resources. AI agents, while promising, must handle these constraints to deliver systemic improvements.
