AI clinical-trial patient recruitment

01Overview and value chain#

Markers EC: AI-based matching of patients to clinical trials and data-driven trial site selection | OECD: Biopharmaceutical clinical development | Regulator: FDA (US), EMA (EU)

AI clinical-trial patient recruitment platforms scan structured and unstructured electronic health record data to identify patients meeting a trial’s eligibility criteria, replacing hours of manual chart review by clinical-research staff. Deep 6 AI optimizes site selection for clinical trials, letting sponsors base site choices on live population and screening data rather than historical site-performance metrics alone. TriNetX unveiled a conversational AI interface and enhanced API capabilities in January 2026, successfully reducing trial costs and timelines as part of platform enhancements designed to democratize clinical-research analytics. Antidote Technologies’ Antidote.me offers AI clinical-trial matching for physicians on an enterprise pay-for-performance model, connecting patients to research directly through referring clinicians. Mendel.ai’s software aims to save nurses time in pre-screening, and academic research on end-to-end AI-powered trial-recommendation systems (TrialMatchAI, published in Nature Communications 2026) explicitly frames patient recruitment as a major bottleneck calling for scalable, automated solutions — the problem this whole vendor category exists to solve. Paradigm Health’s AI-native solution increased screening capacity by 41% in an August 2026 partnership expanding clinical-trial access with Ochsner Health across the US Gulf South, and the company separately submitted a formal response to FDA’s Real-Time Clinical Trials Initiative RFI informed by biopharma sponsors and providers.

The key directions of AI clinical-trial patient recruitment are:

EHR-based eligibility matching: scanning structured and unstructured patient records to automatically identify candidates meeting a trial’s specific eligibility criteria.
Data-driven site selection: scoring candidate trial sites on live patient-population and screening data rather than relying on a site’s historical enrollment track record alone.
Physician and patient-facing matching interfaces: conversational AI and referral tools that connect physicians and patients directly to relevant open trials.
Screening-capacity expansion at health systems: deploying AI-native screening directly inside health-system workflows to expand the population that can be screened for trial eligibility.

Sectoral value chain#

[EHR data access & structuring] ──> [AI eligibility matching] ──> [Candidate patient shortlist]
                                                                    │
                                                        (Site-selection scoring)
                                                                    │
                    [Patient enrollment] <──── [Physician/patient outreach] <─── [Referring-clinician connection]

Fig. 1— Sectoral value chain

Value chain levels#

Level	Description	Key inputs/outputs
EHR data access & structuring	Accessing and structuring a health system’s structured and unstructured patient record data for analysis.	In: Raw EHR data. Out: Structured, analyzable patient data.
AI eligibility matching	Applying AI models to identify patients matching a specific trial’s eligibility criteria.	In: Structured patient data, trial eligibility criteria. Out: Matched candidate list.
Candidate patient shortlist	Compiling the matched candidates into a shortlist for clinical-research staff review.	In: Matched candidate list. Out: Reviewed candidate shortlist.
Site-selection scoring	Scoring candidate trial sites on live population and screening data to inform sponsor site selection.	In: Aggregated population/screening data. Out: Site-selection score.
Referring-clinician connection	Connecting a matched patient to the trial through their referring physician.	In: Reviewed candidate shortlist. Out: Physician referral to trial.
Patient enrollment	Completing the enrollment process for the referred, eligible patient.	In: Physician referral to trial. Out: Enrolled trial participant.

Table 1— Value chain levels

Cross-cutting technologies of the sector:

AI patient-trial matching: machine-learning software that scans structured and unstructured EHR data to identify patients meeting a trial’s eligibility criteria, replacing manual chart review.
Real-time trial site-selection analytics: analytics that score candidate trial sites on live patient-population and screening data, letting a sponsor pick sites likely to enroll quickly.

02US#

The United States hosts the leading AI patient-recruitment vendors, several with direct 2026 health-system deployments and platform enhancements.

Deep 6 AI’s site-selection optimization, TriNetX’s conversational AI, Paradigm Health’s 41% screening-capacity gain#

Deep 6 AI: optimizes clinical-trial site selection using live population and screening data rather than historical site performance alone.
TriNetX: unveiled a conversational AI interface and enhanced API capabilities in January 2026, reducing trial costs and timelines as part of a broader effort to democratize clinical-research analytics.
Antidote Technologies: its Antidote.me platform offers AI clinical-trial matching for physicians on an enterprise pay-for-performance model.
Mendel.ai: its software aims to save nurses time in pre-screening, applying AI directly to the recruitment bottleneck that academic research has identified as the field’s central constraint.
Paradigm Health: an August 2026 partnership with Ochsner Health increased screening capacity by 41% across the US Gulf South, and the company formally responded to FDA’s Real-Time Clinical Trials Initiative RFI.

03CN#

China is covered qualitatively rather than by a live-screened Chinese vendor: candidate Chinese AI patient-recruitment firms searched during this screen returned only generic market-research reports and academic commentary, not a confirmed named vendor, so no Chinese company is tabled below.

No confirmed named domestic vendor; market-research coverage describes the category rather than naming a leader#

Market coverage without a named leader: Chinese-language sources describe the AI clinical-trial patient-recruitment market’s growth and trends extensively, but this screen did not surface a specific China-headquartered vendor confirmed by name in a live 2026 source.
Domestic gap: no China-headquartered AI patient-recruitment platform confirmed by name was found during this screen.

04EU#

No EU-headquartered vendor cleared the confirmation bar on this screen distinct from the US-headquartered platforms already tabled; European trial sponsors and health systems are served by the same global vendors listed above.

Served by the same global platforms rather than a distinct EU-headquartered specialist#

EU deployment of global platforms: the vendors tabled above serve EU clinical-trial sponsors and health systems directly rather than through a confirmed distinct EU-headquartered competitor.
No distinct EU-headquartered vendor confirmed: this screen did not surface an EU-headquartered AI patient-recruitment specialist distinct from the platforms already listed.

05Leading companies and research institutes#

Company / Institute	Country	Key products / platforms	Tech features	Status 2026
Deep 6 AI	🇺🇸 USA	Site Selection platform	Live population/screening-data-based site scoring	commercial
TriNetX	🇺🇸 USA	Conversational AI interface, enhanced APIs	Reduced trial costs and timelines (2026)	commercial
Antidote	🇺🇸 USA	Antidote.me	Physician-facing AI matching, pay-for-performance	commercial
Mendel.ai	🇺🇸 USA	Mendel.ai pre-screening software	Nurse pre-screening time savings	commercial
Paradigm Health	🇺🇸 USA	AI-native screening platform	41% screening-capacity increase at Ochsner Health	commercial

Table 2— Leading companies and research institutes

06Tech stack and innovations#

The AI clinical-trial patient recruitment technology stack combines EHR-scale data access with an increasingly conversational and health-system-embedded AI layer:

Large-scale EHR data matching:
- Deep 6 AI and Mendel.ai apply AI directly to structured and unstructured EHR data to identify eligible patients without manual chart review.
Conversational AI and API expansion:
- TriNetX’s January 2026 conversational AI interface and expanded APIs show the category moving toward more accessible, integration-friendly tooling.
Health-system-embedded screening deployment:
- Paradigm Health’s direct partnership with a regional health system (Ochsner Health) demonstrates AI screening deployed inside clinical workflows rather than as a standalone research tool, delivering a measured 41% capacity gain.

07Value chains and production pipelines#

Industrial pipeline for AI-driven patient recruitment#

┌───────────────────────────┐      ┌───────────────────────────┐
│ 1. EHR data access &        │ ───> │ 2. AI eligibility           │
│    structuring                  │      │    matching                      │
└───────────────────────────┘      └───────────────────────────┘
                                                 │
                                                 ▼
┌───────────────────────────┐      ┌───────────────────────────┐
│ 4. Site-selection scoring   │ <─── │ 3. Candidate patient        │
│                                  │      │    shortlist                      │
└───────────────────────────┘      └───────────────────────────┘
              │
              ▼
┌───────────────────────────┐      ┌───────────────────────────┐
│ 5. Referring-clinician       │ ───> │ 6. Patient enrollment       │
│    connection                   │      │                                   │
└───────────────────────────┘      └───────────────────────────┘

Fig. 2— Industrial pipeline for AI-driven patient recruitment

Stage 1: EHR data access and structuring

The platform accesses and structures a health system’s structured and unstructured patient record data for AI analysis.

Stage 2: AI eligibility matching

AI models identify patients matching a specific trial’s eligibility criteria from the structured data.

Stage 3: Candidate patient shortlist

Matched candidates are compiled into a shortlist for clinical-research staff review.

Stage 4: Site-selection scoring

Candidate trial sites are scored on live population and screening data to inform sponsor site-selection decisions.

Stage 5: Referring-clinician connection

A matched, eligible patient is connected to the trial through their referring physician.

Stage 6: Patient enrollment

The enrollment process is completed for the referred, eligible patient, closing the loop from EHR data to enrolled participant.

Supplier	Region & tags
Deep 6 AI	US
TriNetX	US
Antidote	US
Mendel.ai	US
Paradigm Health	US

AI Recommendation

Key directions:

EHR-based eligibility matching — scanning structured and unstructured patient records to identify candidates meeting a trial’s eligibility criteria.
Data-driven site selection — scoring candidate trial sites on live population and screening data rather than historical enrollment track record alone.
Physician and patient-facing matching interfaces — conversational AI and referral tools connecting physicians and patients directly to relevant trials.
Screening-capacity expansion at health systems — deploying AI-native screening directly inside health-system workflows.

Regulatory:

Patient-matching software touches protected health information, so every vendor in this category operates under HIPAA (US) or equivalent data-protection rules even though the software itself isn’t FDA/EMA-regulated as a medical device.
Recruitment speed is a recognized clinical-development risk factor regulators track indirectly – a trial that can’t enroll fails regardless of drug efficacy, which is part of why FDA runs a dedicated Real-Time Clinical Trials Initiative that vendors like Paradigm Health respond to directly.
Site-selection analytics reduce enrollment-timeline risk in a way sponsors increasingly disclose to regulators and investors as part of trial-feasibility planning.

Companies not in table: this screen confirmed all five drafted US candidates on the first pass – no company was dropped for lack of confirmation.

Processing note: no Chinese or EU-headquartered vendor confirmed by name cleared the bar on this screen – Chinese-language sources described market growth and trends extensively but never named a specific domestic company, and no EU-headquartered competitor distinct from the US platforms surfaced either.

Category boundary: this is distinct from general clinical-trial site-management or CRO services – the buyer here specifically needs AI-driven identification of eligible patients from EHR data and site-selection scoring, not the broader operational management of a trial once patients are already enrolled.