Unlearn Platform / PROCOVA is an AI platform for clinical trial design built around digital twins. A digital twin is a forecast of one participant’s longitudinal clinical outcomes, based on baseline data.
Sponsors simulate trials before protocol lock and apply PROCOVA (Prognostic Covariate Adjustment) to randomized trials. They can also build external or built-in control arms for single-arm studies.
The platform covers three stages: design before protocol finalization, execution with interim analyses, and post-readout re-evaluation. Unlearn’s Solution Scientists work directly with sponsor teams.
The EMA has qualified PROCOVA for Phase 2 and 3 trials with continuous outcomes. Unlearn states the method aligns with FDA guidance. Listed customers include AbbVie, ProJenX, and Acumen Pharmaceuticals.
Pricing
Unlearn does not publish plan names, prices, or subscription terms on its website. Prospective clients contact the company directly to discuss partnership and pricing. The reviewed sources list no self-service sign-up or usage-based pricing.
* Disclaimer: Please note that pricing information may not be up to date. For the most accurate and current pricing details, refer to the official website.
Key Features
-
✓
AI-generated digital twins forecast individual participant outcomes
-
✓
Trial simulations test eligibility, endpoints, and sample sizes
-
✓
PROCOVA adjustment reduces sample size or adds power
-
✓
External and hybrid control arms for single-arm studies
-
✓
Precedent review organizes prior trials into sourced benchmarks
Use Cases
Smaller Randomized Trial Designs
Biostatisticians apply PROCOVA during design to lower the required enrollment. A re-analysis of AbbVie’s Phase 2 Alzheimer’s study showed a 17% smaller participant count.
Open-Label Study Comparators
Sponsors use digital twins as a simulated placebo arm when randomization is not possible. ProJenX applies this approach in its Phase 1 ALS trial of prosetin.
Pressure-Testing Protocols Before Lock
Clinical operations teams simulate designs against inclusion and exclusion criteria. This helps them catch trade-offs and avoid protocol amendments.
Sharper Interim Go/No-Go Decisions
Statisticians add pre-specified digital twin analyses to the statistical analysis plan before unblinding. The result tightens interim estimates and supports stopping decisions.
Re-Reading Borderline Trial Results
Researchers re-analyze completed trials to separate expected disease progression from treatment effect. Clearer evidence can then inform the next study design.
Strengths & Weaknesses
Strengths
The EMA qualified PROCOVA for Phase 2 and 3 trials with continuous outcomes.
A retrospective AbbVie analysis showed a 17% smaller participant count.
One vendor supports design, interim execution, and post-readout analysis.
Unlearn won Fierce’s 2026 AI Innovation Award for Clinical Trial Design.
Named partners include AbbVie, ProJenX, CHDI Foundation, and QurAlis.
Weaknesses
Pricing is unpublished, so buyers cannot estimate cost before contacting sales.
No self-service access is described, so every engagement starts with a consultation.
EMA qualification covers continuous outcomes, which limits scope for other endpoint types.
Public sources list no EDC or CTMS integrations and no security certifications.
Who Is This For?
Biopharmaceutical sponsors: program leaders who want lower trial costs and shorter timelines.
Biostatistics teams: statisticians who design analysis plans, covariate adjustment models, and interim reviews.
Clinical operations leaders: teams that manage protocol design, recruitment, and regulatory submission strategy.
Emerging biotechs: small and mid-size firms running Phase 1/2 or single-arm studies with limited patients.
Frequently Asked Questions
What is a digital twin in Unlearn’s platform?
It is an AI-generated forecast of one participant’s longitudinal clinical outcomes. The forecast uses that person’s baseline data.
How much does Unlearn cost?
Pricing is not published. You need to contact Unlearn to discuss a partnership and costs.
Have regulators evaluated PROCOVA?
The EMA qualified it for Phase 2 and 3 trials with continuous outcomes. Unlearn says it aligns with FDA guidance, so confirm scope with regulators for each program.
At which trial stages can the platform be applied?
Unlearn supports design before protocol lock, interim analyses during the study, and re-evaluation after readout.
How large are the reported sample size reductions?
Results vary by study. A re-analysis of AbbVie’s Phase 2 Alzheimer’s study showed a 17% smaller participant count.
Which therapeutic areas has Unlearn worked in?
Unlearn lists neurological and chronic indications, including ALS, Alzheimer’s disease, and Huntington’s disease.
Can digital twins replace a placebo arm entirely?
In single-arm studies, twins act as external or built-in controls. In randomized trials, PROCOVA reduces control participants rather than removing the arm.
What does onboarding look like?
You request a consultation, then share historical trial data and protocol goals with Solution Scientists. Together you run simulations before finalizing the protocol and analysis plan.
No public integrations are currently available. The source material does not mention EDC or CTMS connections.