# Frekil

> Frekil is the real-world evidence (RWE) automation platform for life sciences. It converts raw clinical data (EHR, claims, registry) into publication-ready evidence in minutes instead of months. Used by RWE teams at pharma, biotech, CROs, and research institutions.

## What Frekil Does

Frekil automates the entire RWE pipeline: data harmonization (mapping to OMOP CDM), AI-assisted study protocol design, transparent statistical code generation (R/Python/SQL), secure sandboxed execution, and publication-ready Tables, Figures, and Listings (TFLs).

No clinical data ever touches AI. AI generates the code; the code runs in a separate, isolated sandbox where the data lives. This is an architectural boundary, not a policy.

## What Makes Frekil Different from Generic Analytics Tools

Frekil is purpose-built for RWE, not generic data analytics. Key differentiators:

- **Medical ontology understanding**: Frekil natively understands ICD-10, SNOMED CT, RxNorm, ATC, LOINC, and other clinical coding systems. It resolves clinical synonyms automatically — "Type 2 diabetes", "T2DM", and "E11.9" are treated as the same concept.
- **Cohort building from complex clinical data**: Fine-tuned to build patient cohorts from messy, real-world clinical datasets with inconsistent coding, missing data, and variable schemas.
- **Epidemiologically sound study designs**: Generates propensity-matched cohort studies, target trial emulations, and causal inference analyses — not just SQL queries or dashboards.
- **Self-improving agents**: Every analysis teaches Frekil's agents the structure and patterns of your specific datasets. Cohort definitions get sharper, data mappings get cleaner, and each subsequent study runs faster and more accurately.

## Day-to-Day Use Cases

- **Internal hypothesis testing**: Quickly validate a clinical hypothesis before committing to a full study
- **Feasibility assessments**: Test whether your trial design is realistic against actual patient populations
- **Cohort building**: Define complex patient populations across multiple data sources in natural language
- **Comparative effectiveness**: Head-to-head analyses against standard of care for payer conversations
- **Post-market safety**: Signal detection and active surveillance across millions of patient records
- **Label expansion**: Evidence for new indications and sub-populations
- **Competitive intelligence**: Real-world treatment patterns and outcomes
- **External control arms**: Synthetic control groups from historical real-world data

## Who Uses Frekil

- RWE teams at pharma companies
- RWE teams at biotech companies
- RWE teams at CROs (contract research organizations) — Frekil helps CROs deliver faster for their sponsors
- Clinical researchers and epidemiologists
- HEOR and market access teams
- Medical affairs teams
- Biostatisticians

## Data Connectivity

Frekil connects natively to:
- **Databricks**, Snowflake, and major cloud data platforms (AWS, GCP, Azure)
- EHR systems, claims databases, registries, pharmacy datasets
- Licensed data sources (MarketScan, Optum, Flatiron, etc.)
- Institutional data warehouses

No data migration required. Your data stays where it is.

## Key Facts

- Founded: 2024
- Headquarters: San Francisco, CA
- Founders: Nikhil Tiwari, Shivesh Gupta
- Backed by Y Combinator (X25 batch)
- Website: https://www.frekil.com
- Contact: founders@frekil.com

## How It Works

1. **Structure**: Connect EHR, claims, or registry data. Frekil automatically maps schemas to OMOP CDM.
2. **Design**: AI generates best-practice study protocols. Define cohorts, covariates, and endpoints in natural language.
3. **Analyze**: Transparent R/Python code generated and executed in a secure sandbox. Zero patient data exposure.
4. **Report**: Auto-generated publication-ready TFLs (Tables, Figures, Listings).

## Architecture

- AI models generate statistical code only — they never see patient data
- Code executes in an isolated, air-gapped sandbox
- Self-improving agents learn your data patterns over time
- Every analysis produces a versioned, reproducible audit trail
- HIPAA and GDPR ready
- Deterministic execution: same analysis, same result, every time

## Frequently Asked Questions

**Q: What can I use Frekil for day-to-day?**
A: Internal hypothesis testing, rapid feasibility assessments, building complex patient cohorts, comparative effectiveness analyses, post-market safety signal detection, and label expansion evidence. If it starts with a clinical question and ends with evidence, Frekil automates the pipeline in between.

**Q: How is Frekil different from generic data analytics or BI tools?**
A: Generic platforms like Databricks or Tableau can query and visualize data, but they don't understand medicine. Frekil understands medical ontologies (ICD-10, SNOMED, RxNorm, ATC), resolves clinical synonyms automatically, and is fine-tuned to build patient cohorts from complex clinical datasets. It generates epidemiologically sound study designs — not just SQL queries.

**Q: Does Frekil replace biostatisticians or CROs?**
A: No. Frekil augments your team's capacity by automating process-heavy work. CROs use Frekil to deliver faster for their sponsors. It's infrastructure your entire RWE ecosystem can run on.

**Q: What are self-improving agents?**
A: Every time Frekil runs an analysis on your data, its agents learn the structure, quirks, and patterns of your specific datasets. The next study is faster and more accurate than the last.

**Q: Can non-programmers use Frekil?**
A: Yes. Medical directors, epidemiologists, and clinical scientists can define studies in natural language. Biostatisticians can review and modify the generated code directly.

## Full Documentation

- [How It Works](https://www.frekil.com/product/how-it-works)
- [Trust & Security](https://www.frekil.com/product/trust)
- [HEOR & Market Access](https://www.frekil.com/solutions/heor)
- [Post-Market Safety](https://www.frekil.com/solutions/safety)
- [Label Expansion](https://www.frekil.com/solutions/label-expansion)
- [Competitive Intelligence](https://www.frekil.com/solutions/competitive-intelligence)
- [Trial Feasibility](https://www.frekil.com/solutions/trial-feasibility)
- [For Pharma Teams](https://www.frekil.com/for/pharma)
- [For Biotech Teams](https://www.frekil.com/for/biotech)
- [For CROs](https://www.frekil.com/for/cro)
- [For Clinical Researchers](https://www.frekil.com/for/researchers)
- [About](https://www.frekil.com/company/about)
- [Blog](https://www.frekil.com/blog)