Trial design decisions are $10–50M decisions. The intelligence behind them shouldn’t live in unstructured PDFs.
ClinicalTrials.gov is the canonical source of truth for what competitors are doing — but the protocols, SAPs, and supplementary documents are trapped in unstructured PDFs and semi-structured XML. We built the extraction layer to make it queryable.
Registry and protocol extraction
- •Structured extraction from ClinicalTrials.gov XML, protocols, and SAPs
- •Endpoints, comparators, populations, estimand frameworks — all queryable
- •Every value traced to its source document
Competitive landscape databases
- •Structured databases across your therapeutic area
- •Track how competitors design trials, select endpoints, and define populations
- •Queryable by indication, phase, sponsor, or any extracted field
Living intelligence
- •New registrations and protocol amendments captured as they appear
- •Results postings and status changes flagged automatically
- •Your competitive landscape stays current without manual review
The questions your team spends weeks answering manually
Endpoint selection
- •What primary and secondary endpoints are competitors using in your indication?
- •How is endpoint selection trending over time?
- •Which PRO instruments are gaining traction with regulators?
Comparator arm design
- •What are competitors using as active comparators vs. placebo, and in which populations?
- •How are comparator strategies shifting as standard of care evolves?
- •Which comparator choices are HTA bodies rewarding?
Estimand frameworks
- •How are sponsors handling intercurrent events — treatment switching, dropout, rescue medication?
- •What statistical analysis strategies appear in competitor SAPs?
- •How are ICH E9(R1) estimands being implemented in practice?
Subgroup definitions
- •How are competitors defining and stratifying patient populations?
- •Which biomarker-defined subgroups are emerging in your therapeutic area?
- •How do subgroup strategies differ across geographies and regulatory bodies?
Trial design patterns
- •Adaptive designs, basket trials, platform trials — what are competitors adopting?
- •How are dose-finding strategies evolving?
- •Which novel designs are regulators accepting?
Landscape monitoring
- •New trial registrations in your therapeutic area as they appear
- •Protocol amendments and status changes that signal competitor pivots
- •Results postings before they hit the publication cycle
How it works
1
Define your therapeutic area and questionsTell us the indication, the competitors, and what you need to track — endpoints, comparators, estimands, populations. We design the extraction schema together.
2
We build your competitive intelligence databaseAI-powered extraction from ClinicalTrials.gov registry data, protocols, SAPs, and supplementary PDFs. 97% extraction accuracy with human expert oversight at every stage.
3
You receive structured, queryable intelligenceStructured databases you can query by sponsor, indication, phase, endpoint, comparator, or any extracted field. Every value traced to its source document.
4
We keep it currentNew registrations, protocol amendments, status changes, and results postings are captured automatically and integrated into your existing database. Your landscape intelligence never goes stale.
What takes your team weeks of manual protocol review, we extract in days.
97% extraction accuracy across protocols, SAPs, and registry data
Every extracted value traced to its source document and page
Queryable databases, not static reports or spreadsheets
Living intelligence — new registrations and amendments captured automatically
Complete data isolation — HIPAA compliant, no training on your data
Service, not software — we build and maintain the database, not a platform login
Built by researchers, for researchers
Cofounder & CEO, in previous lives he has worked on Stanford NLP's most popular open-source project, researched multi-agent AI, and studied Systems Design Engineering at the University of Waterloo.
