Nxtcure Labs

Running a clinical trial involves a long chain of manual, specialized work: scoring site feasibility, drafting protocols, tracking site activation against benchmarks, monitoring risk indicators, triaging adverse events, resolving data queries, running biostatistics, and maintaining an inspection-ready trial master file. This work is typically distributed across many people and vendors in a traditional CRO model, creating paperwork burden and slow cycle times, for example contract activation benchmarked at 77 days. Sponsors running oncology trials need this work done faster without losing the clinical, safety, regulatory and financial judgment calls that require a qualified human.

Nxtcure Labs' platform, Galen, uses thirteen purpose-built AI agents that run a study end to end in lifecycle order across four pillars: start-up and design (feasibility, protocol, activation), clinical operations (recruitment, monitoring, retention, termination watch, supply), data/safety/analytics (data management, safety, biostatistics), and documentation/quality (medical writing, TMF audit). Every agent action with clinical, safety, regulatory or financial consequence is routed to a single human review queue, where a qualified person (safety physician, medical monitor, principal investigator or regulatory lead) approves or rejects it, with each card decided exactly once and recorded with who and when. The platform logs plain-language reasoning for every action including skips and failures, treats agent crashes as high-risk review tasks rather than silent failures, and is demonstrated running on CDISC SDTM data against a synthetic Phase 3 NSCLC study and the CDISCPILOT01 dataset.

The platform is described as designed to the FDA and EMA's January 2026 guiding principles for AI in drug development, positioning it against incoming regulatory guidance on AI use in trials.