Commentary Contributed by Usama Dar, CluePoints 

March 20, 2026 | In clinical trials, every second counts, both for patient outcomes and sponsor ROI. As data volumes grow and oversight expectations tighten under frameworks like ICH E6(R3), teams are being asked to do more, faster and with greater precision. It is no surprise then that agentic AI, which promises to deliver efficiencies far beyond previous models, is the current hot topic across the industry. Analysis suggests agentic AI could transform eight out of 10 workflows and up to 95% of life science roles may have agentic teammates in the future. 

As with any emerging technology, the risk is not that agentic AI fails. Rather, it just becomes another buzzword. To deliver real value in clinical trials, we must understand its capabilities, practical applications, and the governance guardrails required to ensure we continue to turn artificial intelligence (AI) into human intelligence. 

What is Agentic AI? 

The industry is currently operating in an AI paradox. Eight in 10 companies are now using GenAI in at least one business function. However, more than eight in 10 still report no contribution to their earnings from this investment. While enterprise-wide tools like chatbots have been scaled quickly, function-specific use cases, which have higher potential for direct impact, largely remain stuck in pilot mode. 

Agentic AI is seen as a way to break out of this paradox, shifting AI from a reactive tool to a proactive, goal-driven collaborator that frees up human capability for where it is needed most. Agentic AI framework uses autonomous “agents” to perform tasks, learn from new information, and make decisions within defined boundaries. It combines machine learning, large language models, and natural language processing to move beyond prompt-based interaction toward goal-driven execution. Key features include decision making, problem solving, autonomy, interactivity, and planning. 

Imagine you are running a clinical trial. If you wanted to ensure data collection remains compliant, GenAI might generate a checklist based on historical data and regulatory guidance. In contrast, an agentic system could continuously monitor incoming data, flag emerging compliance risks, and recommend corrective actions before those risks escalate. 

Data Management Applications 

One area that has the potential to benefit from agentic AI is clinical data flow. Modelling suggests the use of multiple agents, working alongside human reviewers and biostatisticians, could boost productivity in data programming and management by 60%. 

If we look specifically at medical coding, traditional manual coding was time-consuming, inefficient, and error prone. By combining agentic AI with risk-based quality management (RBQM), organizations can focus on critical-to-quality factors and align with evolving regulatory expectations under ICH E6(R3) and quality-by-design principles. For example, instead of manually coding adverse events and concomitant medicines, high-confidence terms can be coded automatically, and synonyms can be mapped to standardized WHO drug terms. At the same time, proactive alerts can flag only high-risk or unusual drug combinations and dosage forms and indications can be used for contextual coding.  

The benefits extend beyond efficiency: reduced staff burden, continuous learning that improves consistency over time, and the ability to address quality risks before they impact patients or timelines. 

Strategies for Implementation  

Delivering on the possibilities of agentic AI requires more than technology. It needs organizational transformation that aligns people, processes, and platforms for regulatory-ready, data-driven clinical trials.  

Just as we saw with RBQM implementation, effective change management, training, and a compelling narrative about the benefits of change will be vital. We should consider where we can build on existing successes in areas like medical coding, medical and safety review, and query detection. 

It will also be essential to establish clear governance frameworks, robust audit trails, and transparent measurement standards to ensure traceability, explainability, and regulatory readiness. We need to understand what is working — and what isn’t — if we are to transform resistance into adoption and, ultimately, enthusiasm. 

Finally, while agentic AI introduces new possibilities for clinical trial data management and RBQM, it is not a replacement for human expertise. It is an amplifier of it. The future of clinical oversight will not be defined by more dashboards or more alerts. It will be defined by intelligent systems that surface what matters so human experts can act with greater confidence, clarity, and speed. 

Usama Dar joined CluePoints in April 2025 and serves as chief product and technology officer. He holds a master’s degree in computer science from Hamdard University (Karachi, PK) and a master’s degree from the University of Leicester (UK). He brings over 20 years of experience leading technology innovation across healthcare, publishing, and e-commerce. Usama has held senior roles at Elsevier, Huawei, and Sciensus, where he led large-scale platform transformations, cloud modernization, and AI adoption in highly regulated environments. At Westwing, he helped build the company’s original tech stack and later returned to scale the business with modern SaaS architecture. He can be reached at usama.dar@cluepoints.com.