Trial delivery rebuilt on agentic AI.
We run trials. We do not consult on them. Two years of operating agentic AI on live studies gives us a different view of what is possible — and what is not.
We are operators, not advisors.
Most firms that talk about AI in clinical trials are describing what they have evaluated. We are describing what we have run. That distinction matters because the failure modes of agentic systems only become visible under real operational pressure — not in a demo environment.
Our trial delivery practice covers the full study lifecycle: protocol development, site selection and activation, data management, regulatory submissions, and closeout. AI is not a layer we add on top. It is the operating infrastructure.
What we deliver
Decentralized clinical trials (DCT)
Remote patient enrollment, electronic consent, wearable data integration, and virtual site visits. We have run DCT designs across multiple therapeutic areas and know where the regulatory bodies draw the line between acceptable and not.
Digital health technologies (DHT)
Sensor-based endpoints, patient-reported outcomes via app, and continuous monitoring. We select DHTs based on regulatory acceptance history, not vendor relationships.
Agentic AI operations
Protocol-to-CRF generation, automated data review, real-time query management, and CDISC-compliant dataset production. Human review at every stage. No black-box outputs.
Regulatory strategy
FDA pre-IND meetings, IND submissions, Type B and C meeting requests, and NDA/BLA support. We write the documents and attend the meetings.
Biostatistics and data management
SAP development, SDTM and ADaM dataset production, TLF generation, and integrated summary of safety and efficacy. Agentic systems handle the serialization; statisticians handle the science.
Site network and patient recruitment
A curated site network built over two decades, with performance data on enrollment rates, protocol deviation rates, and data quality. We do not use site lists. We use site relationships.
DCT is not a feature. It is a design decision.
Decentralized elements reduce patient burden and expand geographic reach, but they introduce new operational complexity: remote monitoring, data integrity across devices, and regulatory acceptance of novel endpoints. We have navigated all of it. We know which elements FDA will accept without a meeting and which require one.
Remote enrollment and eConsent across 13+ countries
Wearable and sensor endpoint validation
Virtual site visit protocols accepted by FDA
Hybrid designs balancing site and remote elements
What sponsors receive
| Phase | Deliverable | Indicative timeline |
|---|---|---|
| Study start-up | Protocol, IB, ICF, site contracts, IRB submissions | Weeks 1–12 |
| Site activation | Site qualification, training, system access, first patient in | Weeks 8–20 |
| Data management | CRF design, edit checks, data review, query management | Ongoing |
| Regulatory | IND maintenance, safety reporting, agency correspondence | Ongoing |
| Biostatistics | SAP, SDTM/ADaM datasets, TLFs, CSR statistical sections | Weeks 40–60 |
| Closeout | Database lock, final datasets, TMF archival, site closeout | Weeks 52–72 |
Running a trial or planning one?
We work with sponsors at any stage — from protocol concept to NDA submission. The first conversation is always with David.