Australian BPO Tackles HR Documents

BACKGROUND

A need to make historical data digital

Advance Record Management is one of the leaders in providing document storage, cataloguing, archiving and retrieval in Australia. In early 2020 they were asked by a Victorian agency to find a solution that would take their services one step further. They needed to process over 1 million loose-filed human resource records/documents into a hierarchy that would be searchable within the client’s content management system.

As standard policy, the client retains employee records for a minimum of twenty years. This meant that Advance Record Management would have to find a solution for the records of approximately 23,000 past and present employees, in 500 boxes of paper files from a variety of locations

THE CHALLENGE

Hardcopy to Electronic

Finding a solution that could not only scan, but read a variety of different types of forms and formats, such as payroll, contracts, general correspondence, leave forms and performance reviews was a challenge. This solution also needed to ensure these documents could be readily found in the content management system when required.

The other challenge was that approximately 8,000 of these records needed to be done within tight frames due to staff transitions and transfers.

THE SOLUTION

Enter TCG Process’ DocProStar: a platform specifically designed to intelligently automate document-centric business processes through a capture, understand and act infrastructure.

DocProStar is a unified, web-services-based platform incorporating best-of-breed artificial intelligence (AI), learning bots and validation rules, and is extensible with technologies like robotic process automation (RPA), Google Vision and others. This ensures a higher degree of efficiency in documentdriven processes.

For Advance, DocProStar was able to classify, identify and extract data from unstructured documents in a variety of formats, with a high accuracy, eliminating the concern for human error.

To ensure Australian data security requirements were met, DocProStar was hosted on local servers.

The entire project took less than 7 weeks, including configuring the classification workflow for all the documents within DocProStar, and configuring the output to be indexed and exported automatically into the content management system.

THE RESULT

Speed, accuracy and a very happy client
DocProStar took 7-10 days to process 1 million documents, including scanning them into readable PDFs then classifying, OCRing, indexing and exporting all documents into the client’s content management system, by employee name and in chronological order.

Automation saved an estimated 50% reduction in man hours over the course of the project