Unlearn

Unlearn provides an AI-based software platform for clinical development, organized around what the company describes as the sequence of decisions that shape a clinical trial. The platform spans three areas: Plan (trial planning and simulation), Monitor (advanced trial monitoring) and Analyze (trial analyses using AI-generated digital twins). Planning tools include Scout, which searches, structures and summarizes literature and regulatory precedent from sources such as PubMed, ClinicalTrials.gov and drugs@FDA; Hindsight, which exposes harmonized clinical trial and real-world datasets for validating clinical and statistical assumptions, population characteristics and endpoint benchmarks; and SimLab, which builds and compares reproducible trial-design scenarios covering endpoints, inclusion/exclusion criteria, sample size and constraints.

The core technology generates digital twins of trial participants that forecast each participant's control-arm outcomes at future time points, which the company applies to reduce control-arm size and increase statistical power. Advanced Trial Monitoring converts blinded trial signals into curated insights intended to surface execution and data-quality issues during a study. Unlearn states it works in therapeutic areas including neuroscience, immunology and metabolic disease, with published work in Alzheimer's disease, ALS and Huntington's disease.

Business model

Unlearn sells AI software and analytics for clinical development to pharmaceutical and biotechnology sponsors, engaging through collaborations and case-study partnerships tied to specific studies and therapeutic programs [0].

Traction

The company reports approximate real-world results of a 33% control-arm size reduction and more than four months of enrollment time saved, a 20% sample-size reduction achieved (28% potential with Bayesian designs plus digital twins) and about $250K per-patient program savings. A retrospective reanalysis of the ADCS DHA Phase 3 Alzheimer's trial (402 patients, 51 sites) is cited as flagging ADAS-Cog anomalies roughly two months into enrollment, about seven months before formal DSMB review. Digital twins are reported to have increased power from 80% to 90% on ADAS-Cog11 at 18 months using PROCOVA, and a 33% control-arm reduction is cited for the Phase 3 bapineuzumab dataset. Multiple biopharma collaborations were announced between November 2025 and July 2026 [0].