Building an Insurance Market Intelligence Platform
French insurance data startup
A French startup needed to build a comprehensive database of insurance and brokerage companies. We delivered the full data stack: scraping agents, third-party API integrations, a central database, and a query API. The founder focused on sales and product while we handled everything data.
The Challenge
The client was building a product that depended entirely on data they didn’t have yet. They needed a comprehensive, structured database of insurance and brokerage firms, pulled from scattered, unstructured sources across the web. As a startup, they had no data engineering team and a tight budget.
The product’s value depends on completeness and accuracy, but the data doesn’t exist in one place. It’s spread across regulatory registries, professional directories, company websites, and third-party APIs, each with different formats and update frequencies. The client had a web developer for the front end but no capacity to build the data pipeline.
Our Approach
We built and operated the full data pipeline. Phase one (4 weeks) focused on scraping agents for the highest-value sources, handling anti-bot protections and unstructured data. Phase two (12 weeks) added the central database, API connectors, additional scraping agents, and the query API that powers the client’s front end.
What We Built
Web scraping agents
Custom agents collecting data from regulatory sites, professional directories, and company websites. Built to handle anti-scraping protections and unstructured page layouts.
Third-party API integrations
Connectors to external data providers, enriching the scraped data with additional company attributes and market signals.
Central database
A structured, deduplicated database of insurance market actors, designed for fast querying and continuous updates from all collection sources.
Query API
A production API that exposes the database to the client’s front-end application, supporting the search and filtering their customers rely on.
Results
The client launched their product on a complete, production-grade data backend without hiring a data engineering team. The founder focused entirely on commercial development and the front-end product. The engagement expanded as new data sources emerged, and Stratalis continues to operate and extend the platform.
Before & After
| Metric | Before | After |
|---|---|---|
| Data engineering team | None | Full pipeline, outsourced |
| Data sources | Scattered, manual | Automated, multi-source |
| Time to market | Blocked on data | 18 weeks to production |
| Founder's time on data | Primary bottleneck | Zero |