IQVIA vs Unlearn.AI (2026)
IQVIA sells clinical research services, health data and technology to drug companies, and much of its AI arrives inside those relationships. IQVIA.ai, launched in March 2026 with NVIDIA, is a catalog of agents for literature review, deep research on IQVIA's own Med-R1 model, field force work, market insight and cohort building. In October 2026 IQVIA added Life Science Models that predict trial outcomes. The company reports more than 150 agents deployed and 19 of the top 20 pharmaceutical companies starting to use them. Unlearn sells one method: digital twins that forecast how each trial participant would have progressed on control. In a randomized trial the forecast becomes a covariate under PROCOVA, which the European Medicines Agency qualified in 2022 for phase 2 and 3 primary analysis. Sponsors work with Unlearn's scientists to shrink the control arm or lift power. Unlearn now also offers twins as comparators in single arm studies. IQVIA brings breadth: data, people and agents across a sponsor's clinical and commercial work. Unlearn brings depth on one statistical question, how few patients a trial can enroll and still get its answer. The two can sit side by side, since one runs the operation and the other changes the design. A sponsor consolidating AI with the partner that already runs its studies and data leans toward IQVIA. A sponsor facing an expensive or slow trial in a disease Unlearn models leans toward Unlearn.
- IQVIA.ai offers a catalog of agents across literature review, deep research, field force, market insight and cohort building.
- The Med-R1 Deep Research Agent runs on IQVIA's own Med-R1 8B medical reasoning model.
- Agents can deploy in IQVIA's technology stack or the client's, and connect to IQVIA data such as MIDAS and real world cohorts.
- IQVIA Life Science Models predict trial outcomes, with 85 percent of predictions on target or clinically close in a first test across 50 trials.
- The agents come with IQVIA's research services, data and consulting in one relationship.
- Digital twins forecast each participant's outcome on control from baseline data.
- The PROCOVA analysis holds a 2022 qualification opinion from the European Medicines Agency for phase 2 and 3 primary analysis.
- In a randomized trial the method keeps the false positive rate controlled even when a forecast is wrong.
- Reanalyses of completed phase 3 trials showed a 33 percent smaller control arm and more than four months of enrollment saved.
- Custom generators can be built and run entirely inside the sponsor's cloud.
This comparison is published by AI Health Index, an independent research platform that compares healthcare AI vendors objectively. IQVIA and Unlearn.AI are each graded against the same capability taxonomy, from each vendor's own public materials and the regulatory record, under the AI Health Index verification standard. No vendor pays for placement, and no vendor has reviewed this page. How this evidence is graded
Plain facts
| Fact | IQVIA | Unlearn.AI |
|---|---|---|
| Primary category | Clinical Trials AI | Clinical Trials AI |
| Founded | Not recorded | 2017 |
| Headquarters | Research Triangle Park, North Carolina | San Francisco, California, United States |
| Website | iqvia.com | unlearn.ai |
Side by Side
Select any grade to read the note behind it.
Product updates
The latest logged changes from each vendor, newest first, each linked to its source.
IQVIA
-
Model / architecture
IQVIA launched IQVIA Life Science Models, a suite of purpose built AI models on NVIDIA Nemotron and AWS infrastructure.
Source -
Product / capability
IQVIA launched Predictive Clinical Development, a suite covering site selection and study design, an automated start up path it calls Push Button Start-Up, and real time data cleaning aimed at shortening the gap to database lock.
Source
Each vendor in one paragraph
Written to be quoted whole. Each paragraph states what the AI Health Index verified about the vendor, with the caveats attached.
IQVIA sells clinical research services, health data and technology to life sciences companies, and IQVIA.ai, launched in March 2026 with NVIDIA, packages its AI as a catalog of agents for literature review, deep research, field force, market insight and cohort work. The agents run on IQVIA's own Med-R1 model and NVIDIA's Nemotron stack, can deploy in IQVIA's or the client's technology, and reach 19 of the top 20 pharmaceutical companies by IQVIA's count. Life Science Models, added in October 2026, predict trial outcomes. No agent's limits or outcome study are published, and IQVIA publishes no prices.
Source: AI Health Index, October 2026
Unlearn is a San Francisco company whose Digital Twin Generators forecast how each trial participant would fare on control, sold to drug and biotech sponsors as a partnership with its statisticians. The forecasts enter a randomized trial's analysis through PROCOVA, a method the European Medicines Agency's CHMP qualified in 2022 for continuous outcomes, so a sponsor can enroll fewer control patients for the same power. Generators exist for 26 conditions, with published support in Alzheimer's disease, ALS and Huntington's disease. The evidence comes from reanalyses of completed trials, the single arm comparator use goes beyond what regulators qualified, and pricing is not published.
Source: AI Health Index, October 2026
Questions buyers ask
How does IQVIA compare to Unlearn?
IQVIA and Unlearn both sell AI to drug sponsors, at very different scales. IQVIA offers a catalog of agents through IQVIA.ai, alongside its research services and health data, covering literature review, deep research, commercial work and trial outcome prediction. Unlearn offers one method, digital twins of trial participants, which let a randomized trial reach the same answer with fewer control patients.
IQVIA makes sense for a sponsor consolidating AI with its main research partner, and Unlearn for one trying to shrink a specific trial. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
How much do IQVIA and Unlearn cost?
Neither IQVIA nor Unlearn publishes a price. IQVIA's agents are sold inside enterprise agreements that usually bundle data, services and clinical research, and the IQVIA.ai catalog lists availability but no terms. Unlearn sells partnerships by study, program or portfolio and points to savings instead, such as about $2 million on one phase 1/2 study and a 33 percent smaller control arm in reanalyzed phase 3 trials. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
Can IQVIA or Unlearn make a clinical trial smaller or faster?
Unlearn's digital twins are built for exactly that. In reanalyses of completed trials they cut 92 participants from a phase 3 ALS study and shrank the control arm by 33 percent in phase 3 trials from a top five sponsor. IQVIA works on speed through operations, with Predictive Clinical Development for site selection, study design and startup, which it says can save up to two years over a typical ten year program, and Life Science Models that forecast trial outcomes.
Neither has shown the gain prospectively in a completed registrational trial. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
What regulatory standing do IQVIA and Unlearn have?
Unlearn's PROCOVA method holds a 2022 qualification opinion from the European Medicines Agency's CHMP for primary analysis of phase 2 and 3 trials with continuous outcomes. It covers the method rather than a specific generator, and not the newer single arm use. IQVIA makes no device claim for its agents and publishes no position on how they sit inside validated GxP systems or how a model change is handled, though its research services operate under those rules. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
How do IQVIA and Unlearn protect sponsor and patient data?
IQVIA's annual report says its global data centers and IT controls are covered by an annual SOC 2 Type II audit, without saying whether IQVIA.ai falls inside it. IQVIA states no business associate role for the agents and does not say whether customer feedback trains shared models.
Unlearn announced SOC 2 Type II in 2024 without publishing the report, commits to HIPAA safeguards where it signs a business associate agreement, and lets sponsors build generators entirely inside their own cloud. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified October 25, 2026. No vendor pays for placement.
Related comparisons
Other published head to head assessments involving these vendors or their closest peers. The full set for this category is on the Clinical Trials AI page.
IQVIA is based in Research Triangle Park, North Carolina, and files annual reports with the SEC. Unlearn is based in San Francisco. Neither publishes a price. IQVIA acts toward sponsor data as a service provider, data holder and software vendor at once, and does not say which role governs each agent, what its agents may do without a person, or whether customer prompts and feedback train models other customers use.
Unlearn does not publish where a twin based single arm comparator stops being appropriate, or how its generators perform across age, sex, race or disease stage. Its efficiency figures measure statistical power and sample size, not patient outcomes.