The key to the future: Intelligent document recognition

But the market is in transition to IDR, led by automated invoice processing and explanations of benefits (EOBs). Both of those consist of many forms that are expensive to process using trained manual labor but can be automated with appropriate sets of rules. Interfacing with and gaining maximum benefit from those applications requires an understanding of the backend process needs.

In the case of invoice processing, the highest value is gained when integrated with order-to-pay or procure-to-pay software. The suppliers of that software have developed solutions in conjunction with enterprise resource planning (ERP) offerings to help users better manage their supply chain. In the case of EOBs, the software is being sold in conjunction with benefit administration systems installed in hospitals or other providers.

In each case, the key is using the software to understand the document layouts allowing users to manage their business process more effectively. Capture has expanded from simply scanning paper into images for use in archives and workflow ECM to, for example, scanning batches of mortgage applications to ensure that all the documents are present and in the right order. It has expanded from scanning and capturing data used in claims or tax systems, to becoming the efficient entry point for unstructured data into business processes.

By being able to understand and intelligently tag unstructured data, IDR capture opens a whole new set of markets in the BI and BPM space, as well as ECM. Because it relies on rules-based understanding, it also changes and expands the channels of distribution much more into vertical markets based on the rule sets. This may mean that traditional suppliers of capture software will be challenged by vendors of BPM and BI, but we tend to think that those companies will be content to license capture software. It fits with the traditional IT approach toward capture and data entry forms processing (i.e. this is not our area, please give us the data to process).

However, those who understand and control the business rules will be those who prosper, resulting in changes to the distribution channel to vertical solution suppliers. Those resellers who cling to capture as just a front end for ECM or DM systems will fade into obscurity.

Concurrent with this, mailroom automation is starting to be implemented. The software, which is designed to convert physical mail into e-mail and automatically route it, was introduced a few years ago. At that time, it was being encouraged after the anthrax scare in the United States because Congress decided that it would no longer accept physical mail. But mailroom automation on a more general level has not been well accepted. We believe this is because of:

• availability of low-cost mailroom staff,
• extensive use of lockboxes to presort mail,
• questions about privacy, and
• questions on how to deal with mass mail.

Mailroom automation techniques have been adopted in the United States even though pure mailroom has not been implemented. One example in traditional medical claims is that exceptions must be out-sorted manually. Mailroom processing techniques allow that out-sorting to be handled automatically from the images.

Meanwhile, in Europe, pure mailroom automation processing seems to be increasingly accepted. National post offices are being denationalized and are looking for ways to increase revenues while keeping costs under control. It is a huge challenge for an industry built on picking up physical mail and delivering it. Led by Denmark and Germany, electronic mail delivery seems to be catching on. The user must sign a waiver of confidentiality and any mass/junk mail is delivered by hand much more slowly.

Corporations are also starting to install the systems, and Pitney Bowes, the giant of mail output processing, announced that it will offer a service to scan the mail to its corporate clients at CeBIT. This will begin, I was told, in Germany and then the United Kingdom, followed by the United States.

Conversion service is big business in the United States. Service bureaus scan and index documents for backend ECM systems. Clearinghouses convert paper medical claims into EDI. Service companies convert e-mail into faxes. It seems that some form of mailroom processing will become popular here in a few years time.

Capture is set for continued double-digit growth over the next few years even though business is increasingly being transacted electronically, and nicely structured e-forms are reducing the amount of paper forms being generated. The fact is that business will never eliminate unstructured, untagged data even though transactional paper seems to be in decline.

Capture systems that are now being developed have the ability to interrogate that unstructured business data and add the metadata that the recipient company or organization can understand. Based on this need, we are forecasting a compounded annual growth rate of 16 percent through 2009. Capture is an interesting business for those interested in KM.