Why image hasn't become part of the process at most banks.

PAPER PLAYS AN overwhelming role in banking. Banks' externally originated correspondence, files, reports, and payment documents probably total some 60 billion to 70 billion pieces of paper each year.

Although the value of these many bits of paper is high, the cost is also high and the interest in converting to the "paperless office" is almost eternal.

Image processing is the only technology that can replace this stream of paper on a one-to-one basis. By scanning paper documents, processing interactions can be transferred to the computerized images.

This technology is unique because encoded data is not generally created or, in many cases, even wanted. Once indexed and filed, the images exist primarily to be read or accessed by bank workers.

Image technology is not new. In the late 1970s, it was used in remittance processing. By 1982, Banc One was experimenting with the capture of payee information to be returned on Merrill Lynch cash management account statements. Today, checking statements returned with images are not even leading edge.

It is accepted that image technology has the potential to cut the cost of item processing data entry.

For example, Ohio's Huntington Bancshares Inc. uses Unisys' check courtesy amount read algorithm to read 41% of its checks without manual data entry with an error rate of only 0.9%.

Image technology is also used relatively routinely in wholesale lockbox, customer correspondence, consumer credit application processing, signature verification, and mortgage custody applications.

It will soon be incorporated in automated teller machines, to help customers see what they have deposited. And it may even capture signatures at point-of-sale terminals to help stem the rising tide of chargebacks.

Yet, many doubts remain. The fact remains that image technology is far from prevalent. Most industry desks do not have image-equipped terminals or access to software applications that process images.

The cost benefits have really not been proven and the amount of paper flowing in and out of banks continues to increase. No industry-specific application packages yet employ image technology. No third-party processors offer image services.

As this year's American Banker/Ernst & Young technology survey showed, implementations of image technology consistently lag expectations.

Three factors have contributed to slower levels of image implementation.

First, extensive reengineering or redesign of today's workflows is often not anticipated, so the conversion period is longer, slower, and less beneficial than planned.

Second, banks have looked for value-added applications to make images produce customer revenues in addition to cutting back-office costs. Yet, success has been rare.

And finally, generally unforeseen implementation and technical problems have arisen.

Bank workflows are almost always determined by the flow of paper. To replace the paper while maintaining the workflow is like paving the cowpath.

Instead of a linear progression of work from one workstation to the next, imaged work can be broken up, stored, and accessed only when and where needed.

For example, doing item processing with imaging requires a very thorough workflow redesign. Encoding is done at the end -- instead of the beginning -- of the proof process. On-us checks do not have to be encoded.

The order in which items are sent to the amount-entry function does not depend at all on the order of the items themselves. Various software algorithms can now assist in balancing blocks of items that previously could only be done manually.

In file folder applications, prioritizations and short-term work queue adjustments require new job descriptions and new clerical organization.

If proper controls are present, parallel actions can be taken by geographically dispersed production workers, speeding completion. Realtime tracking and measurement of workers can lead to productivity pay and/or realtime matching of workflows with job skills.

Workflow software that helps index, monitor, and retrieve images -- and control the rate at which all employees work on them -- is becoming an integral part of all banking image applications.

It is easy to be overly optimistic about workflow redesign. The best new workflows may not be discernible until well into the implementation process.

In the beginning, costs and time may actually go up due to conversion and adjustment problems. Improving productivity may depend upon extraneous events, like redesigning a bank's forms so that they support the new workflows.

In short, there is a generalized learning curve that the industry must follow to make effective use of image technology. It's not as easy as it looks.

Image technology is a back-office cost-cutting tool. But the technology will progress further and faster when images can contribute directly to revenue production. Many ideas have been proposed, but progress in this area to date has been very slow.

Some ideas we've heard include:

* Linking corporate account analysis customers to the image system to get back bulk or online transmissions of check images.

* Restructuring retail checking accounts to offer three tiers of service and price: no checks returned; images of checks returned; the real checks returned.

* Addition of data to retailers' proprietary payment documents that would then be captured with the image technology and sold back to the retailer.

* Credit card chargeback systems that capture images of paper sales drafts to resolve chargeback disputes without requiring direct merchant effort.

To date, however, the aggregate payoff from these ideas has been low. No outright sales of images to customers have occurred. Many wholesale and corporate customers are themselves not yet invested in this technology. Image interchange capabilities are still ahead.

Better customer service is a real enough front-office benefit, but is difficult to quantify. Moreover, too many of the potential ideas come from the payment system, while most banking image applications are of the back-office file folder type. While it very likely that additional value-added applications will be found, in the interim banks will continue to move more slowly than they would otherwise.

Unexpected technical issues are a fact of life. There are no standards yet for exchanging images within or between banks. Mainframe-based image systems may be at variance with some banks' plans for a more open, distributed approach. Some workflow software is relatively new and not suitable for banks' high volume needs. Especially in the payment system, the documents themselves (checks, deposit slips, cash-out tickets) are quite old in design and are simply unsuitable for an image-technology dominated world.

For example, there are no standards for checks. Yet if there were, the speed of implementing imaged item processing would be improved substantially, along with the benefits and the ease of conversion. This one step could do more than anything else to make check image technology work better.

Some image applications remain expensive, especially those that require extensive transmission facilities. When very high levels of low-cost bandwidth become available, the industry may see a real uptick in the speed at which geographically dispersed branches or back-offices begin to make use of this technology.

The bottom line is that image moves more slowly than generally expected because paper displacement is such a new way to do business. Paper is literally ingrained in the way we think about everything. For example, most courts accept only a paper document as legal proof This is only now beginning to change.

We are literally suffering from failure of the imagination -- failure to really understand what to do to take advantage of this innovative approach.

We will undoubtedly improve in this area and image will yet become a commonplace technology on bankers' desks. Yet until that happens, bankers should be careful and search for only applications offering the highest payoff. As always, a thorough understanding of the business, the current workflows, and the technology will be one of the keys to future success.

For reprint and licensing requests for this article, click here.
MORE FROM AMERICAN BANKER