Turning Hidden Document Data Into Value That Can be Mined

0915CanonArtToday’s businesses are facing many challenges as technology innovation continues to transform how and where business is conducted. One of the biggest challenges is integrating technology within business workflows. For instance, with the adoption of mobile technology, work is being conducted from multiple locations. The ability to conduct business anywhere and anytime is clearly changing user behavior, and has a significant impact on the workflow processes that have been designed based on paper and human interaction rather than mobile and digital.

At the same time, businesses are being pressured to reduce costs, be more efficient and increase revenue. To improve efficiencies, businesses are deploying solutions designed to increase workflow speed and overcome the delays and costs associated with paper workflows. However, although organizations invest in technologies to automate workflows, they still struggle with costly processes that contain bottlenecks caused by manual data entry and antiquated workflows.

Driving business process improvements and workflow automation requires executive leadership, resources and expertise. But business process improvement often has no clear home and often leads to fragmented approaches and mixed results. Moreover, there is competition between the drive for business innovation and the drive for business process efficiency.

Most organizations focus on workflow improvements through incremental workflow modification rather than through innovation. Step-by-step improvements are easier to plan, more predictable, faster to implement and more achievable. Improvement through innovation is risky, requiring bold new ideas and approaches as well as new technologies or skills the organization may not have.

Document data entry is prime for workflow automation and innovation

One area that is currently prime for improvement through both process change and innovation is the conversion of document-resident data into ERP formatted data. Almost all business processes, from the front office to the back office, require paper documents to complete the process and bridge the digital systems (see Figure 1). High-profile processes include accounts payable invoice entry, logistics and freight, and many others. The lesser-known processes include real estate property management, pharmaceutical trials and many more.


What is the innovation opportunity?

Currently, a significant part of the BPO workflow taken offshore more than 10 years ago is still completed using manual data entry. At the time, lower-cost, plentiful offshore resources made sense even if double-key entry was used. But in the past 10 to 15 years the data capture and extraction technology at home has improved beyond comparison. In this setting the innovation is the change necessary to move away from manual data-entry offshore to a modern Optical Character Recognition (OCR) and associated business rules and workflow technology at the point of document capture. Customers are awakening to the benefits of applying intelligence and quality control at the document first-point-of-touch to eliminate errors and apply computer processing power on the data being captured. This was not possible until recent years. Even more powerful is the idea of capturing far more data than previously allowed by economic constraints, for the purpose of extracting information and knowledge for insights, analysis, and better decision making. For example, in the past it was customary to limit the number of data fields entered manually in order to contain costs — just the approximately six to eight fields necessary to complete the transaction were entered. Now, using automated capture, additional fields are captured, matched, and validated using business rules and the data is routed or distributed using workflow — all leading to a faster and a better business.

What is the Automation opportunity?

The automation opportunity for moving document-based data transcription into digital systems consists of leveraging best-in-class OCR, data extraction, business rules engines, matching, routing and automated classification. Other technologies such as barcode recognition may also be incorporated depending on the document and the process.

According to a recent IDC study sponsored by Canon Business Process Services, the No. 1 workflow automation initiative for businesses is capturing paper content.  See Figure 2.


Automated data conversion applications can read any document and media type (paper, PDF, spreadsheet, fax, etc.) where data needs converting to a standard ERP format. The process involves two steps: conversion from arriving media formats (such as paper, fax, and email attachments) and validation processing based on business rules.

The processing staff only works the exceptions that business rules cannot resolve and that require human intervention. For example, using an invoice as the document, it is possible for upwards of 98 percent of the invoices to pass through the process hands-free. We have a client receiving 12,000 paper invoices a month and only about 30 require manual intervention in the workflow.

Calculating your ROI using automated capture and workflow

When making the cost/benefit calculation for document data conversion, the focus has to shift from the cost of the activity to the benefits resulting from the activity. In addition, comparing the cost of the old data transcription to the new automated data extraction has to be made based on an apples-to-apples basis because the automated method stretches into the process and eliminates processing work.

Using the invoice as a document example, one can calculate the cost of the invoice entry as the sum of scanning plus the cost of labor for data entry. The result is just a transfer of the invoice data into the ERP system. Under an automated process the exact transfer of data is performed by the automated workflow. However, in the same time span the automated process can perform a number of additional tasks that in the manual process are completed later at additional cost. These tasks include matching the invoice data against a number of data points such as the vendor master, previously received invoices, PO line items and approval table. And, workflow actions can instantly be taken to accept, reject or route the invoice.

These tasks are hard to itemize and price separately and therefore hard to compare to a manual process. Moreover, how do you place a value on time saved and insight gained from the data? In the invoice example it is possible for the automated scan, data extraction and validation to move the invoice from receipt to ready-to-pay stage in a matter of minutes. The manual process may take days. The automated process can flag duplicate invoices and new suppliers, and provide the rich data needed to analyze the purchases and make decisions about supplier pricing or item volumes, etc. This is the value of capturing that extra data, made possible by an automated workflow.



As organizations battle disruptive forces beyond their control on one hand, they must also battle paper intensive legacy processing on the other. The technologies and off-shoring approaches used to improve document workflows 15 years ago are outdated and ineffective by today’s standards. Today we have technology that can capture the document, extract its data and process the transaction without human intervention. But change is required to make this work. To make the new process pay off in terms of efficiency and ROI, specialized document processing centers are usually the answer. These centers can serve as the hub for data conversion, bridging business systems in order to provide better access to information and enable the organization to mine the data for information and insight.

This article originally appeared in the September 2015 issue of Workflow.

Ted Ardelean is Director of Research & Development Marketing for Canon Business Process Services. Visit www.cbps.canon.com for more information.