> ## Content Index
> Fetch the complete content index at: https://acentricresearch.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# AI in clinical trials: Junction Brief – November 2026
- URL: https://acentricresearch.com/research/ai-clinical-trials-brief-2026-11/
- Published: 2026-11-20T22:40:27.000Z
- Updated: 2026-11-21T09:39:37.000Z
- Author: Kevin Meyer
- Tags: ai-biotechnology, junction-brief

> Acentric Research is produced with AI tools under human review. The authors are not financial professionals, and nothing here is investment advice or a recommendation to buy or sell any security. AI analysis can be wrong; check every claim against its cited source.

This Brief is an initial scan, not a full investigation. The method and scores behind it are in [AI × Biotechnology in the Landscape Report](https://acentricresearch.com/research/landscape-2026-11/#ai-biotechnology).

### Overview

ARPA-H, the U.S. Advanced Research Projects Agency for Health, states that clinical development of drugs and biologics often takes more than a decade, costs $1–2 billion and fails 90% of the time. \[1\] Software that learns from past data is being proposed as a way to shorten some of those steps. This Brief covers that area: artificial intelligence (AI) used to design trials, match patients to them and run them.

What stands out is how mixed the evidence is. A federal agency has opened a funding call aimed at much shorter trials, and a European regulator has qualified a statistical method that a trade publication describes as AI-driven. Companies have announced financings for AI recruitment and patient-selection tools. But no source opened measures whether AI makes trials faster or more likely to succeed. This Brief describes aims and company claims. The dated sources opened run from September 2022 to December 2025, plus program and company pages read in November 2026; later developments were not searched.

### Current state

**A federal program.** ARPA-H’s SURPASS program seeks to build “perpetual, adaptive platform trials” that combine predictive computational models, shared trial infrastructure, common control groups and real-time analysis. AI is one of several skills the solicitation asks of teams, alongside statistics and trial design, operations and regulation. ARPA-H states a goal of shortening trials from more than a decade to less than four years. Its page lists an open solicitation that was accepting submissions on 2026-11-18\. \[1\] The page describes an aim, not a result.

**A company, in its own words.** Unlearn is one example of a company in this area. Its home page is headed “Streamline Clinical Trials with AI and Digital Twins of Patients”. It defines a digital twin as an “AI-generated forecast of one participant’s clinical outcomes”, and describes using twins to model the patients a trial expects to enroll, to sharpen interim analyses, and to serve as the comparator so that a trial can enroll a smaller control arm. Its page states results from re-analyzing completed trials, for example a 33% smaller control arm in Phase 3 trials from a top-five global sponsor (company claims). \[2\]

**A regulatory step.** A trade publication reported in September 2022 that the European Medicines Agency (EMA) qualified Unlearn’s PROCOVA procedure for Phase 2 and 3 trials with continuous outcomes. The report describes the framework as the first regulatory backing of a machine-learning method for reducing sample sizes in pivotal trials, and says the method could support gains in power, a reduction in sample size, or both. \[3\]

**Companies that raised money.** Pulse 2.0, a news site, reported in September 2025 that Trially had raised $4.7 million for Margo, an AI tool for trial recruitment, and in December 2025 that Valinor had raised $13 million in seed funding for a machine learning platform that selects patients by predicted response to therapy. Both reports draw on company statements. \[4\]\[5\]

**How far it has come.** The most developed element is one statistical method that a regulator has qualified, according to a trade report. Most of the rest is at the stage of funding calls, company claims and early financings.

### Opportunities

ARPA-H’s program would fund teams to build adaptive platform trials if awards follow the solicitation; the page states no award. \[1\] A regulator’s qualification of a method is one step toward wider acceptance, though the reported qualification covers a specific kind of analysis. \[3\] If trial designs that use fewer control patients are accepted more widely, the saving would show up in trial size and enrollment time, which companies such as Unlearn already describe as the benefit. \[2\]

### Challenges

- **Regulation.** A method qualified by one regulator for one kind of analysis does not settle the question for other uses of AI in trials. The reported EMA qualification covers Phase 2 and 3 trials with continuous outcomes. \[3\]
- **Evidence.** The speed and savings claims above are aims or company claims. No independent measurement that AI shortens trials was found in the sources opened.
- **Capital and scale-up.** Two financings were found, seed rounds of $4.7 million and $13 million; financing was not searched systematically. \[4\]\[5\]
- **Finding the work.** ClinicalTrials.gov has no coded field for AI use, so trials that use AI can only be found by text search. \[6\]

### Data sources

- **ClinicalTrials.gov**, for registered trials. It has no coded field for AI use, so trials that use AI have to be found by text search. \[7\]\[6\]
- **EMA qualification opinions**, for methods a regulator has accepted for a stated use. \[8\] FDA publishes guidance on AI in drug development separately; it was not opened for this Brief.
- **ARPA-H program pages** and NIH RePORTER, for federal funding calls and grants. \[9\]\[10\]
- **SEC EDGAR**, for financings and company descriptions by companies that file. \[11\] Private companies’ rounds are mostly known through press releases and news reports.
- **Company pages**, which are the main source for what AI tools do in trials and are company claims. \[2\]

### Changelog

- 2026-11-18: First published.

### Sources

1. ARPA-H, “SURPASS: Simulation-augmented, Real-time Platform Adaptive Seamless Trials” (program page). [arpa-h.gov](https://arpa-h.gov/explore-funding/programs/surpass?ref=acentricresearch.com) (accessed 2026-11-18)
2. Unlearn.ai, Inc., company home page (company claims). [unlearn.ai](https://www.unlearn.ai/?ref=acentricresearch.com) (accessed 2026-11-18)
3. Vishnu Priya N, “EMA qualifies Unlearn’s AI-driven approach for smaller trials”, Clinical Trials Arena, 2022-09-29 (trade publication; includes company statements). [clinicaltrialsarena.com](https://www.clinicaltrialsarena.com/news/ema-qualifies-unlearn-approach/?ref=acentricresearch.com) (accessed 2026-11-18)
4. Amit Chowdhry, “Trially AI: $4.7 Million Raised For AI-Based Clinical Trial Recruitment Platform”, Pulse 2.0, 2025-09-22 (company-sourced report). [pulse2.com](https://pulse2.com/trially-ai-4-7-million-raised-for-ai-based-clinical-trial-recruitment-platform?ref=acentricresearch.com) (accessed 2026-11-18)
5. Amit Chowdhry, “Valinor Raises $13 Million To Advance AI-Driven Patient Selection For Clinical Trials”, Pulse 2.0, 2025-12-13 (company-sourced report). [pulse2.com](https://pulse2.com/valinor-13-million-seed-funding/amp/?ref=acentricresearch.com) (accessed 2026-11-18)
6. Acentric Research, Landscape Report data appendix, evidence.json (Acentric Research’s record of ClinicalTrials.gov: “AI use is not a coded field; would need text search”). [acentricresearch.com](https://acentricresearch.com/report-data/#landscape-report-november-2026) (accessed 2026-11-18)
7. National Library of Medicine, ClinicalTrials.gov. [clinicaltrials.gov](https://clinicaltrials.gov/?ref=acentricresearch.com) (accessed 2026-11-18)
8. European Medicines Agency, “Qualification of novel methodologies for medicine development (QoNM)”. [ema.europa.eu](https://www.ema.europa.eu/en/human-regulatory-overview/research-development/scientific-advice-protocol-assistance/qualification-novel-methodologies-medicine-development-qonm?ref=acentricresearch.com) (accessed 2026-11-18)
9. ARPA-H, “Programs”. [arpa-h.gov](https://arpa-h.gov/explore-funding/programs?ref=acentricresearch.com) (accessed 2026-11-18)
10. National Institutes of Health, “RePORTER”. [reporter.nih.gov](https://reporter.nih.gov/?ref=acentricresearch.com) (accessed 2026-11-18)
11. US Securities and Exchange Commission, “Search Filings” (EDGAR). [sec.gov](https://www.sec.gov/search-filings?ref=acentricresearch.com) (accessed 2026-11-18)