A Regional Property Data Company
How a growing property imaging company unlocked new revenue streams while enabling AI-powered sales tools—without exposing customer PII.
A growing regional roof repair company with a data-rich sales application. Their system aggregates satellite imagery, public records, and proprietary data sources to help sales teams identify and prioritize potential clients based on roof condition and property characteristics.
The company wasn't tech-forward—they'd built a capable sales tool, but the underlying data architecture was tightly coupled with customer PII. Names, addresses, and contact information were intertwined with the property analytics that made the system valuable.
Sales-Focused Team
Limited engineering resources for data infrastructure
Rich Property Data
Satellite imagery, records, and analytics in one API
A third party wanted to license their property data to better understand lumber positioning and market trends. But the API exposed customer names and addresses alongside the valuable roof quality metrics.
They wanted to add AI-powered insights to their sales app—letting reps ask questions about opportunities and patterns. But feeding raw customer data to AI models raised obvious concerns.
Two different use cases. Two different privacy requirements. Same underlying data. Building separate APIs for each would require engineering work they couldn't afford—and maintaining them would compound the problem.
Instead of rebuilding their data infrastructure, they enrolled their existing sales API into DataHarbor and created purpose-specific Virtual APIs.
DataHarbor connected to their existing sales app data source. No data migration. No new infrastructure. The original API continued serving their internal sales team unchanged.
Partner Data View
The partner gets exactly what they need for market analysis—city-level location data and roof quality metrics—without any customer PII. Delivered as a familiar REST API for straightforward integration.
AI Agent View (via MCP)
Tokenizing last names lets the AI understand property ownership patterns—"this token owns three properties in the region"—without ever seeing the actual name. Sales reps get AI insights about multi-property owners without the model accessing real identities.
Marketplace Data Product
With the privacy controls already proven on their partnership and AI Virtual APIs, the team realized they could safely position their property data for broader monetization. They created a fourth Virtual API with even tighter parameters and published it to the DataHarbor Global Marketplace—opening a new revenue stream they hadn't originally planned for.
Months later, an update to their sales application started exposing customer phone numbers in the data API—a field that had never existed before.
DataHarbor's Response
Fathom, DataHarbor's schema monitoring system, immediately flagged the newphoneNumberkey as an uncovered field requiring attention.
Instead of phone numbers silently flowing to partners and AI systems, the team received an alert and made deliberate decisions:
Data licensing deal launched in weeks, not months. Revocable access gave both parties confidence.
Sales team gained AI insights without exposing customer identities to models.
Unexpected revenue stream from publishing governed data to the Global Marketplace.
Schema changes don't slip through. Every new field requires explicit governance.
See how Virtual APIs can help you reuse data safely—across partners, products, and AI.
Join the waitlist and be among the first to experience plug-and-play MCP servers. No hassle, no complex onboarding — just revolutionary API management.