A market leader in technology for the bankruptcy industry required a solution to reduce the manual handling of bankruptcy petition documents. On a yearly basis, this firm manually processes millions of bankruptcy petition documents. Using a manual approach, data entry workers would read through the forms and key the pertinent field information into their internal database. The process was costly and inefficient. This prompted the firm to search for a solution to recognize, capture, and extract the field information automatically from the bankruptcy petition documents. Automating the process was complicated due to the semi-structured nature of the fields within the bankruptcy documents.
This market leader in technology licensed SoftWorks AI’s Trapeze to automatically extract field information from bankruptcy petitions located within schedules A, B, and C. The firm purchased a license to process 10,000,000 pages per year. SoftWorks AI’s automated data extraction solution replaced the manual solution. The extraction from Schedules A, B, and C are in full production. Automation rates for Schedules A, B, and C are above 90% accuracy, with fewer than 10% of the documents requiring human intervention. Based on the initial success of Phase I of the project, automated data extraction from Schedules D, E, and F are in pre-production.
The costly and time-consuming manual approach was replaced by a fully automated solution. This solution achieves accuracy rates above 90%. In these cases, a human does not need to be a part of the process. The reduction of human interaction from the process has produced measurable results related to overall speed, accuracy, and cost savings. With SoftWorks AI’s data extraction solution, the bankruptcy petition documents are handled more accurately through machine recognition, at lower overhead cost, while greatly accelerating the entire workflow process.
|Reduce costs associated with handling bankruptcy petition documents.||By implementing SoftWorks AI’s Trapeze automated data extraction solution, data entry personnel and manual handling costs were drastically reduced.|
|Improve overall accuracy of forms processing.||The automated solution combines machine recognition with human recognition to achieve accuracy rates superior to the manual approach.|
|Accelerate processing time of document handling.||With the automated solution, the client is able to process more documents in less time.|
|Extract data from Schedules A (Real Property), B (Personal Property), C (Property Claimed as Exempt) of bankruptcy petitions.||Schedules A, B, C are in full production. Due to the success of the first phase, Schedules D, E, F are in pre-production.|
|Generate XML based output of the extracted field items for connection to proprietary backend system.||The extracted information is output into XML format, which is directly uploaded into the client’s system.|
|Provide a log file for exception reporting.||For the 5-10% of documents that are incomplete or inaccurately recognized, a log file is generated for further handling.|