642
Total Corporate Bankruptcy Filings in 2023
40.4%
Business Bankruptcy Filings Rate Increase in 2023
18,926
U.S. Business Bankruptcy Claims Filed in 2023
EXECUTIVE SUMMARY
TensorIoT's AI-powered document processing solution automated Inforuptcy's ingestion pipeline, significantly reducing processing time and enabling rapid data delivery to customers.
GOAL
How can we automate the document ingestion process to efficiently process the backlog and streamline data extraction?
RESULTS
The TensorIoT solution transformed the time-consuming manual process into an efficient automated system, reducing document processing time from days to hours. This enabled rapid processing of the extensive backlog and allowed Inforuptcy to deliver valuable information to their customers swiftly and effectively.
Background
Inforuptcy, a financial services company with the largest database of bankruptcy assets, provides a service built around processing hundreds of thousands of forms and sharing this information with potential buyers. However, their manual, human-dependent document ingestion pipeline hindered their ability to process the extensive backlog of 700,000 documents efficiently.
The Challenge
Inforuptcy receives thousands of bankruptcy PDF forms and had a backlog of 700,000 pages. They needed a solution to extract specific data from the ingested forms and feed it into their database, eliminating the need for manual reading and data collection. The time-consuming manual process led to delays in providing valuable information to their customers.
The Solution
TensorIoT developed a solution that automated the document ingestion process. The system uploaded backlogged documents to an Amazon S3 bucket, triggering a Lambda function that passed the PDF images to Amazon Textract (OCR). Textract processed, extracted, parsed, and formatted the text and data from the scanned documents, subsequently storing the structured information in Inforuptcy's database.