Competitor Price Collection for a Major Insurer
Major French insurance company
We automated competitor insurance quote collection across auto, home, motorcycle, and boat products. The client went from 1,500 manually entered profiles to 28,000 automated profiles, cutting data entry by 95% and halving the collection cycle.
The Challenge
The client’s pricing and actuarial teams were manually filling out competitor quote forms and recording the results. The process was too slow, limited in volume, and consumed team time that should have gone to analysis.
To stay competitive, the pricing team needed to benchmark against competitors regularly. This meant visiting competitor websites and comparison platforms, filling out quote simulation forms for different customer profiles, and recording results. Done manually, the team processed about 1,500 profiles per cycle, taking six months. Volumes were insufficient for meaningful analysis, and the teams spent their time on data entry instead of pricing strategy.
Our Approach
We built a robust automation pipeline that fills out insurance quote forms programmatically across competitor sites and comparison platforms. The system handles anti-scraping protections including multi-factor authentication via email and SMS. The project was delivered in three phases: auto (15,000 profiles), home (10,000 profiles), then motorcycle and boat (3,000 profiles).
What We Built
Form automation at scale
Automated completion of insurance quote simulation forms across competitor sites and comparison platforms.
Multi-product coverage
Auto, home, motorcycle, and boat insurance. Each product line with its own profile configurations and target sites.
Anti-bot handling
Built to handle aggressive protections including email and SMS-based multi-factor authentication.
Phased delivery
Auto first (15,000 profiles, 8 weeks), then home (10,000 profiles, 4 weeks), then motorcycle and boat (3,000 profiles, 2 weeks).
Results
The pricing team’s time shifted from data entry to actual analysis. With dramatically higher profile volumes and faster collection cycles, the team can now benchmark with statistical confidence across every product line. The operational model scales without additional headcount, and first results were visible within 12 weeks.
Before & After
| Metric | Before | After |
|---|---|---|
| Profiles per cycle | 1,500 | 28,000 |
| Manual data entry | Hours per week | 95% less |
| Collection cycle | 6 months | 3 months |
| Error rate | Manual entry errors | < 1% |
| Time to first results | --- | 12 weeks |