Frekil

Frekil provides an AI-powered Real-World Evidence (RWE) infrastructure that automates the epidemiological workflow from raw clinical data to publication-ready evidence. It uses self-improving AI agents to test hypotheses, build cohorts, and run end-to-end studies while grounding outputs in literature and ensuring data stays within user environments.

Frekil connects to EHR, claims, and registries, then orchestrates an 8-stage pipeline (Literature Review, Question Clarifier, Cohort Builder, Causal DAG Builder, SAP Generator, Data Extractor, SAP Executor, Report Writer) to generate RWE. It generates statistical code in R, Python, or SAS, runs in sandboxed environments, and maintains a strict architectural separation where data never leaves the client’s environment while AI writes the code and executes analyses. Users interact via conversational interfaces, review causal DAGs and cohort definitions, and iteratively refine statistical plans before execution. The platform supports mapping to OMOP CDM, handling biases, and producing outputs such as survival analyses, hazard ratios, and p-values, across use cases like HEOR, post-market safety, label expansion, competitive intelligence, and trial feasibility.

Who it’s for: Pharma/biotech companies, contract research organizations (CROs), HEOR teams, and clinical research groups conducting real-world evidence studies.

Backed by Y Combinator; mentions enterprise security/compliance (HIPAA/GDPR) and live pipeline usage, indicating traction and funding signals.