The Asset Receivership Search Engine

Distressed asset data was fragmented and inaccessible

For real estate investors and legal professionals, identifying properties in receivership requires monitoring an overwhelming number of sources - court filings, legal notices in print newspapers, municipal records, and property registries. These sources don't talk to each other, and the critical data often lives in paper-based formats that resist digital search. The result: high-value opportunities go unnoticed until it's too late.

Cross-domain data that doesn't connect

Solving this required bridging three fundamentally different data worlds:

• Analog print notices: Legal receivership announcements published in newspapers, in varied formats and fonts, with no structured digital equivalent.
• Geospatial context: Property addresses need to be matched to map coordinates, ownership history, zoning data, and tax records from multiple municipal databases.
• Speed requirements: Real estate opportunities move fast - a system that processes last week's notices isn't useful. Near-real-time ingestion was essential.

A multi-modal search architecture

TendersLab built an end-to-end pipeline that transforms print legal notices into a searchable, map-linked intelligence platform.

1. Automated newspaper digitization

Building on our Newspaper Reading System, we deployed a computer vision pipeline (Mask R-CNN + Tesseract OCR) that continuously processes new newspaper editions, isolating legal notice sections and extracting structured data: property addresses, receivership type, court reference numbers, and hearing dates.

2. Entity resolution and geocoding

Extracted addresses are passed through a multi-step entity resolution layer that disambiguates addresses, corrects OCR errors using NLP context, and geocodes each property using the Google Maps API and Israeli cadastral databases. Properties are enriched with zoning, tax history, and ownership data.

3. Interactive map intelligence

The final output is a live map interface where analysts can filter by property type, date, region, or legal status. The platform surfaces neighborhood-level clustering of distressed assets - a signal that often indicates broader economic shifts before they appear in official statistics.

Democratizing access to distressed asset data

• 3x more opportunities: Analysts can now evaluate three times as many potential opportunities per week compared to manual search workflows.
• Hours to minutes: Initial screening time for any property - combining legal status, zoning, and tax history - dropped from hours to under five minutes.
• Market intelligence: Neighborhood-level clustering analysis identifies emerging distress patterns weeks before they surface in conventional market reports.