Classify Litigation Documents and Extract Key Entities for Credit Scoring

Industry Law

Document types Litigation Document

Location China

Language Simplified Chinese

Technology involved NLP, Entity Extraction, Classification, Deep Learning, Knowledge-base, Web Crawling, De-Duplication, and AI Recommendation

Background

A credit rating company uses litigation documents as one of the methods to assess the borrower's creditworthiness and to determine the likelihood of default on a loan. For example, litigation documents, such as court filings, judgments, and liens, can provide information about a borrower's financial history, including any past legal disputes or financial obligations. This information can be incorporated into the borrower's credit report or credit score, which can be used by lenders to make lending decisions.

Challenge

The credit rating company understands that crawling and analysing litigation documents manually is inefficient. Since they are required to crawl and analyse over 1,000,000 litigation documents every day, the volume of documents is too large for handling manually. It can take a significant amount of time and resources to review each document manually, which may not be feasible, especially within a reasonable timeframe. Besides, manual analysis of such a large number of documents is prone to human error. It is challenging for individuals to maintain a high level of accuracy and attention to detail over a prolonged period of time, especially when faced with a large volume of information. Therefore, they were looking for a solution to automate the analysis process.

Solutions

We developed a cloud-based AI system to support the credit rating company, in mainland China to process over 1,000,000 litigation documents in real-time every day since 2016. The system crawls litigation documents from over 3,000 websites and uses Natural Language Processing (NLP) and Machine Learning (ML) technology to classify and extract useful text information from those non-structural and free style documents. At the same time, the system is needed to crosscheck and verify the information from different data sources, and combine and de-duplicate those repetitive information. Our own-build AI system is efficient enough to process such tremendous amount of documents in real-time while achieving 95% accuracy.

Other Use Cases

Cross-Industry

AI knowledge management with intelligent search
‍Coming Soon
‍

Document ClassificationEntity ExtractionAI Recommendation

Law

AI for intelligent search and decision making
‍Coming Soon
‍

Document ClassificationEntity ExtractionAI Recommendation

Law

Transform the agreement review process with AI
‍Coming Soon
‍

Document ClassificationEntity Extraction

Classify Litigation Documents and Extract Key Entities for Credit Scoring

Background

Challenge

Solutions

Other Use Cases

Start optimizing your workflow and boost efficiency by decluttering your inbox