A use-case prioritization framework for Gen AI in clinical development FIGURE 1 Digital twins for virtual trialsProtocol digitalization ImpactFeasibility Priority use cases Other use casesHigh LowLow High Regulatory submission Recruitment, retention, dropoutData analysisTrial feasibility and site selectionClinical trial design Note: List is not exhaustive. Source: ZS analysis Interviewees placed the five development processes into three prioritization groups: –The Holy Grail – clinical trial design: Optimizing clinical trial design using Gen AI can cut clinical development time and costs significantly. Measuring return on investment (ROI), however, will be challenging, given the impracticality of running comparator trials – with one based on traditional methods and another designed using predictive AI. Feasibility is further challenged by the fact that trial design is the most complex and interconnected use case, necessitating intricate decision-making processes. –The low-hanging fruit – regulatory submission: Writing regulatory submissions has traditionally been a manual process, lasting weeks to months and requiring hundreds to thousands of hours throughout the clinical development cycle.6 Automating this process can save a significant amount of time and money. Accuracy will be paramount, so the model will need to be tuned correctly. –Evolution, not revolution – patient recruitment, retention and dropout prevention; data analysis; and clinical operations: Predictive AI has already brought heightened efficiency to trial feasibility, site selection, clinical operations and data analysis. Gen AI will amplify traditional algorithms, marking a step forward in sophistication rather than a disruptive revolution. Interview subjects identified regulatory submission and clinical trial design as the best places to start. These use cases represent a mixture of a quick pay-off (regulatory submission) and high return on investment (clinical trial design). But to realize AI and Gen AI’s full potential for revolutionizing clinical development, pharmaceutical companies must evolve all five phases. Intelligent Clinical Trials: Using Generative AI to Fast-Track Therapeutic Innovations 6

Intelligent Clinical Trials 2024