BEGIN:VCALENDAR VERSION:2.0 PRODID:-//hacksw/handcal//NONSGML v1.0//EN METHOD:PUBLISH BEGIN:VEVENT DTSTAMP:20260310T223433Z DESCRIPTION:Click for Latest Location Information: http://edw2018.dataversi ty.net/sessionPop.cfm?confid=121&proposalid=9461\n

One of the top 10 fina ncial institutions was having trouble with time-consuming and error-prone m anual data entry tasks. Skilled workers were being allocated towards unskil led work, and offshore contractors had to be hired to process documents bei ng sent to them from customers in a variety of formats and quality. They re quired the elimination of the manual process of extracting data and client information from various insurance documents, arriving daily in mass quanti ty, and automating the process of updating their databases using the extrac ted and validated info. Our team at MorganFranklin Consulting helped to cre ate and document a proof of concept for a new automated workflow using cust om-built Python code that leveraged the latest computer vision techniques a nd performed object detection, document classification, data ingestion, ana lysis, validation, and storage. Reporting metrics were also built out to as sess data extraction accuracy and human capital time savings. This automati on process resulted in saving about 80 FTE hours per month for that specifi c line of business. Topics covered by this presentation include:

\n\n Challenges and complexity associated with automating a process that is appl ied to data originating from different sources with different formats and q uality\n The final solution that is a combination of few approaches and methods\n Clearly articulated and measurable business value\n\n DTSTART:20180426T104500 SUMMARY:Intelligent Extraction and Process Automation DTEND:20180426T114459 LOCATION: See Description END:VEVENT END:VCALENDAR