Secured NPL Data Acquisition Agent

Reads 130 million pages of loan and property files and extracts 330 datapoints per case for due diligence.

ClassifyExtractSearch & answerNPL & credit servicing

The business process

Management of credit portfolios and real-estate assets from non-performing loans, supporting legal due diligence and recovery strategy through massive data acquisition.

What the agent reads

  • Mortgage contracts and mortgage registrations
  • Guarantees and sale orders
  • Court appraisals (CTU), property registers and statements of liabilities

What it produces

  • 330 datapoints including creditor data, property owners and complete cadastral data: sheet, parcel, sub-unit, cadastral income
  • A semantic search engine over the whole unstructured dataset

How it works

Automatic classification and entity/object extraction based on semantic context and layout. An integrated semantic search engine lets analysts retrieve information with natural-language, regex or keyword queries across the entire archive.

Runs on

  • Domain-tuned SLM/LLM/VLM serving
  • Semantic knowledge graph
  • GraphRAG & traversal queries
  • Human-in-the-loop review

Bring the AI to your people, not your people to the AI.