Problem

Frekil addresses the bottleneck between the large volume of real-world clinical data that life-sciences companies possess and the publication- or regulatory-grade evidence they need from it. A single real-world-evidence study can cost $100,000–$1 million and take 6–18 months, with roughly 80% of the effort consumed by data cleaning, schema mapping, formatting, cohort construction, and other “data plumbing” before analysis begins. The killer use case is enabling pharma companies and CROs to test more hypotheses and run studies such as safety monitoring, label-expansion analysis, trial-feasibility work, and external-control-arm research without waiting months for manual workflows or expensive service delivery.

Product / Service

Frekil is an AI-powered real-world-evidence infrastructure platform delivered against a customer’s existing clinical data. It connects to EHR, claims, registry, pharmacy, and licensed datasets in the customer’s own environment, maps messy records to medical standards such as OMOP, and lets medical or epidemiology teams describe a study in natural language. Its agents propose the study design, build cohorts, generate protocols and statistical code in R, Python, or SAS, execute the analysis in a sandboxed environment, and produce reproducible, publication-ready outputs while keeping patient data outside the AI models.

The benefit is a shift from months-long, handoff-heavy RWE projects to automated analyses that can run in minutes or days, with biostatisticians retaining control over scientific decisions and generated code. Frekil markets the product to both pharmaceutical organizations and CROs, offers per-study and platform pricing, and says it can support production studies within the first week once data is connected.

Market

Frekil competes in the real-world data and real-world evidence software market, particularly the emerging category of AI-enabled evidence-generation infrastructure for pharmaceutical companies, CROs, and commercial or strategy teams. Its competitive set includes established RWE platforms and services such as Aetion, Clarify Health, ConcertAI, IQVIA Real World Solutions, TriNetX, Veeva Vault RWD, Flatiron, and Castor, as well as generic infrastructure such as Databricks and Tableau that customers might otherwise use for querying and visualization. Frekil’s differentiation is its purpose-built understanding of clinical ontologies, epidemiologic workflows, cohort construction, and statistical execution rather than merely offering general-purpose analytics.

Frekil appears to be an early commercial company rather than clearly pre-revenue. Public evidence identifies it as YC P25/X25 and says it was founded in 2025; third-party directories report approximately $500,000 of funding and conflicting revenue estimates of roughly $256,665 to $330,000. Those figures are not independently verified company disclosures and may include the company’s earlier medical-imaging annotation phase, so current RWE revenue is uncertain. The current website provides anonymized user testimonials and a demo-led sales motion, but the reviewed materials do not name customers or disclose a verified customer count.