Unlearn.ai is an artificial intelligence AI biotechnology company based in San Francisco, California, founded in 2017 to advance digital twins of clinical trial participants to build smart control arms and make drug development faster and more efficient.
A well-funded biotech AI leader with strong investor backing and a highly innovative digital twin technology for clinical trials which shows high credibility as an accelerator for drug development.
The company was started by four individuals; Charles Fisher, Aaron Smith, Graham Siegel, and Jonathan Walsh in San Francisco to develop digital twins for clinical trials using AI.
In its first round of funding, the company raised $650,000 led by DCVC to get started and develop the companies first AI models.
In its second round of funding, the company raised $4.2 million led by DCVC Bio to validate the company's digital twin technology.
In its third round of funding, the company raised $12 million from 8VC to grow the team and enhance the companies ability to use AI for clinical trials.
Cloud-based AI infrastructure
Proprietary AI models generating digital twins with longitudinal clinical outcome prediction capabilities for trial control arms
Inferred from product description and industry standards for clinical AI platforms
| Service | Cost | Details | Source |
|---|---|---|---|
| TrialPioneer | Free | Web-based application for clinical trial planning, available to pharmaceutical and biotech enterprises. Provides insights into trial design choices, sample size needs, statistical power, and enrollment timelines. | Unlearn official announcement |
| TwinRCT (Digital Twin Services) | Custom quote | Core product suite of statistical techniques for clinical trials. Available through partnerships with pharmaceutical companies. Milestone-based pricing tied to drug development progress. | Company documentation |
| Full-Service Support | Custom quote | Comprehensive support including AI-powered digital twin generation, trial design optimization, and regulatory pathway guidance. | — |
| Infrastructure Integration | Custom quote | Integration of AI into client's own infrastructure for flexibility in implementation approach. | — |
| Feature | Unlearn.AI | Traditional CROs | Real-World Data Companies |
|---|---|---|---|
| Digital Twin Technology | Yes | No | No |
| AI-Generated Control Arm | Yes | No | No |
| Regulatory Approval Pathway | Yes (EMA validated) | — | Partial |
| Reduces Patient Enrollment | Yes (up to 15%) | No | Limited |
| Trial Duration Reduction | Yes | No | Partial |
| Pricing Model | Custom/Milestone-based | Per-service fees | Custom |
| Free Planning Tool | Yes (TrialPioneer) | No | No |
| FDA/EMA Discussion Status | 2+ years engagement | Standard review | Emerging |
| Primary Focus | Clinical trials acceleration | Trial execution | Data integration |
While both Unlearn and CROs utilize historical data to enhance their respective trial processes, there is a major difference in how they do this - Unlearn utilizes AI powered digital twins to reduce the number of patients required to enroll in trials, whereas CROs execute trials as planned. Unlearn seeks to optimize the upstream aspects of trial design; CROs seek to optimize the downstream aspects of trial execution. The main differences between Unlearn and CROs include: Unlearn provides the ability to perform faster, less expensive trials; CROs provide operational support for trials, but have no advantages when it comes to reducing the size of populations required for trials.
While Unlearn’s TrialPioneer application will be used as a complimentary service to contracting with CROs, pharmaceutical companies may utilize the TrialPioneer application to develop their own trial design before partnering with CROs to execute the clinical trial.
While both Unlearn and RWD companies utilize historical data to improve trials, the two companies differ greatly in how they use historical data - Unlearn's digital twins mathematically validate that no bias exists when generating control arms, whereas RWD companies face regulatory uncertainty regarding the use of historical data for late stage trials. Unlearn has received approval from the European Medicines Agency (EMA) to use its digital twins for the purpose of generating control arms for clinical trials, whereas many RWD solutions have not established clear regulatory pathways for the use of historical data in late stage clinical trials. Additionally, Unlearn focuses primarily on optimizing control arm reductions, whereas RWD focuses on broadening the scope of data used in trials.
As the first mover in the AI driven clinical trial optimization area, Unlearn has been validated by regulatory bodies such as the European Medicines Agency (EMA) as a valid method of optimizing clinical trials.
Unlike traditional statistical tools, which require manual analysis and interpretation of data, Unlearn utilizes AI to automate the process of creating control arms for clinical trials. In addition to being automated, the process of generating control arms utilized by Unlearn is considered to be at a higher level than traditional statistical tools, as Unlearn is focused on optimizing the overall trial design process, rather than providing foundation level tooling like traditional statistical tools. By utilizing Unlearn, users eliminate the need to develop extensive plans for statistical analysis, whereas traditional statistical tools require advanced knowledge and understanding of statistics. In order to ensure the text sounds like a human wrote it, I will rewrite each numbered item between the marker BEGIN_TEXT and END_TEXT below. I will NOT change the date information, nor will I add or remove any fact-based information. Please let me know if you would like me to do so.
The Unlearn TrialPioneer application simplifies the process of developing a clinical trial design. However, in order to perform advanced statistical analyses on the clinical trial data collected during the clinical trial, statistical tools will still need to be utilized.
The free TrialPioneer application from Unlearn’s TrialPioneer application allows all to utilize free trial design optimization tools, while the costs associated with hiring an expert are retained by consultants. Consultants rely upon their own expertise when providing quantitative design scenarios for trials, whereas Unlearn provides the same design scenarios using computers. Unlearn designs trials in multiple locations; consultants provide trial design services based upon one project at a time.
Compared to the time and expense required to develop a clinical trial plan utilizing a consulting firm, Unlearn’s TrialPioneer application utilizes less time and money.
It creates artificial intelligence AI generated digital twin representations of each clinical trial participant. These digital twins generate predictions of every clinical outcome that will occur at every future point in time, which enables predictive analysis without having to run an actual trial.
It uses AI to search and synthesize relevant scientific and regulatory precedent from PubMed, ClinicalTrials.gov, and drugs@FDA. It organizes and summarizes the findings to enable alignment of your team around the evidence supporting your clinical trial design within days versus weeks.
It analyzes both harmonized clinical trial and real world datasets to evaluate and validate your clinical and statistical assumptions, as well as to determine how your population characteristics, endpoint behavior, and benchmarks can be used to create informed early design decisions.
It automatically generates links, builds, and compares different trial design scenarios based upon your chosen endpoints, eligibility criteria, and sample sizes. Every scenario is also explainable, reproducible, and supported by historical evidence.
Its native architecture is designed to allow for handling of incomplete longitudinal records. Using probabilistic methods and correlations across biomarkers and clinical features, it will infer missing values without biasing your trial outcomes.
No single study or one-time analysis can provide the same level of insights and connections as Unlearn's ability to link together evidence, historical data, and simulations into a transparent workflow.
| Use Case | Clinical Challenge | Unlearn.AI Solution | Business Impact |
|---|---|---|---|
| Early Trial Design & Planning | Design decisions are iterative; assumptions evolve; rationale disappears after review meetings | Unified workspace preserving assumptions, evidence, and results across clinical, biostatistics, and regulatory teams with transparent, review-ready workflows | Faster alignment on design decisions before protocol finalization; defensible rationale across governance reviews |
| Endpoint & Eligibility Optimization | Selecting appropriate endpoints and inclusion/exclusion criteria requires extensive manual analysis | Compare multiple trial design scenarios live; visualize how protocol choices change populations and expected outcomes in real-time | Informed trade-offs before protocol finalization; reduced protocol amendments |
| Sample Size & Population Estimation | Uncertain population characteristics and endpoint behavior lead to oversized studies | Historical data exploration and digital twin simulations validate clinical and statistical assumptions grounded in evidence | Right-sized studies; reduced control group sizes; improved trial success likelihood |
| Precision Medicine Development | Limited insight into which patient subgroups benefit from specific treatments | Digital twins predict how patients respond to treatment and identify subgroup-specific treatment efficacy patterns | Unlocks precision medicine opportunities; tailored treatments for specific patient populations |
| Evidence-Based Design Documentation | Manual spreadsheets and fragmented searches prevent clear documentation of design rationale | Single unified workspace anchoring all assumptions to credible evidence with audit trails preserved across iterations | Defensible protocols ready for regulatory submission; faster governance sign-off; reduced rework cycles |
This tool allows you to assess and compare the characteristics of your target population based on historical data to support realistic estimates of patient enrollment and study feasibility.
The clinical trial design simulations include a cost analysis component, allowing you to optimize the footprint of your study and help plan resources.
The digital twins created by Unlearn are able to forecast all possible clinical outcomes at all future time points, allowing you to have early warning signs regarding the potential safety and efficacy of your trial.
Identify optimal study sites and site network strategies based on population characteristics and protocol feasibility
Live comparison and refinement of protocol choices with immediate visibility into population and outcome trade-offs
