A global financial company faced the problem of distributing and delivering high volumes of incoming physical mail. Managing the mailroom delivery process was manual, costly, and inefficient. Mail volume was increasing year to year, and related personnel and storage requirements grew proportionately.
The response time for critical material suffered because mail could only be viewed by one person at a time. This process required employees to physically distribute mail in an ad-hoc, non-repeatable workflow for their teams to review. Time-sensitive mail was not prioritized any differently than regular mail, which subjected the company to late fees and penalties.
The client also had a need to identify document types and properties with thousands of classes and hundreds of thousands of documents. Document formatting and types changed from year to year. The manual process prohibited them from efficiently sorting mail, tracking history, and retrieving information quickly. Based on these inefficiencies, they sought a solution to automate data capture, reduce costs associated with paper-based mail, and prioritize the classification of mail review.
SoftWorks AI’s Trapeze was purchased to automate and accelerate the distribution process of paper-based mail. A solution was implemented to classify each incoming document from the scanner. After each page is scanned, the software captures the single-page TIFF files produced by the scanner and converts them into multi-page, searchable (OCR), compressed PDF files, or PDF/A for document retention. Each document is then automatically routed to the appropriate employee or group.
At the same time, the system classifies each document, determines the pertinent relevant data, identifies its location on the image, and accurately extracts all key data for detailed analysis. From an end-user perspective, the review process became greatly simplified and faster. Employees now receive their pertinent documents conveniently in their queue, presorted, prioritized and ready to be viewed by any team member on their computer.
The solution yielded several quantifiable results. Automatically routing incoming mail from a scanner enables the organization to reduce costs associated with manually classifying documents. For this automated mailroom classification project, the return on investment was measured in months, not years. In addition to producing a measurable ROI, the entire process was accelerated and a definitive and repeatable workflow was achieved.
To combat the complexity of automatically identifying thousands of document classes, Trapeze utilized its Training Module equipped with machine learning technology. This module was designed to train and test classifiers, report accuracy via email, and automatically back up and place trained binaries (the newly learned code) in the correct location. In order to accommodate document changes, the classifier trained on the latest year’s worth of data. The module also analyzed classifications in order to see how training improved accuracy, and sent an email with a report.
In addition to automation and the classification of digital mail, the solution produced a more efficient image document. PDF conversion is a benefit for both users and the IT department. Converting documents into compressed, searchable PDF files enables users to locate text within paper-based documents instantly. Compression enables users to transmit and access files in a fraction of the time. From an IT perspective, compression generates files with reduced storage size that require less network bandwidth.