A significant part of Unlearn.ai's value is not just its advanced generative models, but its painstaking data harmonization work. The company builds internal machine learning tools to unify complex, disparate data sources like clinical trials and real-world data, which is the essential foundation for creating powerful models.

By using foundation models to analyze vast datasets, companies can create a synthetic 'standard of care' arm for single-arm Phase 1 trials. The AI matches patients based on deep clinical and genomic parameters, providing insights comparable to a much larger Phase 3 trial.

Beyond early discovery, LLMs deliver significant value in clinical trials. They accelerate timelines by automating months of post-trial documentation work. More strategically, they can improve trial success rates by analyzing genomic data to identify patient populations with a higher likelihood of responding to a treatment.

Unlearn.ai's method for late-phase trials (PROCOVA) is acceptable to regulators because it's designed to statistically correct for any bias in the digital twin model. This ensures the model's inaccuracy doesn't affect the trial's final decision procedure or error rate, a critical feature distinguishing it from simply replacing the control arm.

While most focus on AI for drug discovery, Recursion is building an AI stack for clinical development, where 70% of costs lie. By using real-world data to pinpoint patient locations and causal AI to predict responders, they are improving trial enrollment rates by 1.5x. This demonstrates a holistic, end-to-end AI strategy that addresses bottlenecks across the entire value chain, not just the initial stages.

It's impossible to generate human data at the scale of in silico experiments. The key is to create highly accurate simulations of human physiology (digital twins) and then validate their predictions with limited, strategic human data. If the model proves reliable, it could drastically accelerate R&D.

Unlearn.ai strategically avoids diseases where a single biomarker determines progression. Instead, they focus on complex, systematic diseases where many variables each have a small impact on the outcome. These are the areas where sophisticated, multi-variable modeling provides the most significant advantage over standard statistical adjustment.

Unlearn.ai found that scaling digital twins from CNS to oncology isn't about parameter changes. Radically different data structures—like oncology's hierarchy of rare diseases and complex treatment histories—demand entirely new modeling approaches, unlike the more siloed data found in CNS trials.

To de-risk its EMERALD trial for a poorly documented patient population, Resolution Therapeutics first ran a natural history study (OPOL). This provided crucial data to inform the trial protocol and, more importantly, allowed the creation of a matched external control arm, a clever and capital-efficient strategy.