Banks that have installed the first wave of imaging systems for processing deposited checks have seen productivity jump an average of 20%, a recently published study concludes.

The report by Mentis Corp. said 23 banking companies have installed or are testing these "proof-of-deposit" systems, which improve work flows by replacing actual checks with digitized images.

But bankers have been reluctant to divulge the specific productivity gains associated with the adolescent technology.

The Mentis report - which includes data from the 12 banks considered furthest along in adopting check image technology - is perhaps the best independent confirmation yet that vendor promises of double-digit cost savings were not j ust talk.

Bankers say most of their current productivity pins in proof-of-deposit processing result from speeding up key entry of dollar amount information on checks, streamlining the balancing process, and reducing other inefficiencies inherent in paper-intensive check processing.

But they believe even greater gains are possible with continued improvements in optical character-recognition software, which automatically "reads" the amount field on a check. This reduces the need to have operators input data manually.

The Next Step

"We are seeing significant gains in the encoding area" even without character recognition, said Terry Geer, a senior vice president at Huntington Bancshares Inc., Columbus, Ohio. But once the new software is installed, "we expect greater labor reductions within a few months."

Huntington, which has been running all its check volume on an image proof-of-deposit system from Unisys Corp. since the beginning of this year, expects to install character-recognition software within two months.

Although no bank is now running such software in full production, Mentis executives said five institutions are in the late stages of testing.

The results of those tests indicate that image proof-of-deposit systems equipped with character-recognition technology are able to extract payment-amount data from checks at twice the speed of the typical manual key entry operator, who can handle about 1,480 items per hour.

Depending on the accuracy with which the software identifies data - the so-called "read rate," which usually hovers around 50% - the character-recognition feature can boost an image proof-of-deposit system's productivity by 5% to 25%, the Mentis study found.

Reluctance to Jump

Yet despite these numbers, Mentis executives said they do not expect significant numbers of financial institutions to embrace image proof-of-deposit systems before 1994.

Proof-of-deposit processing "is not mission-critical to most banks," said James Moore, president of Mentis Corp., which is based in Eden, Md. "The economic climate is still such that few institutions are willing to take any chances on a project that requires this amount of reengineering."

Mr. Moore said the results from the first installations of International Business Machines Corp.'s newly announced image proof-of-deposit system will significantly affect bankers' decisions on whether to buy the system. It includes character recognition as a standard feature.

Yet it is clear that the multimillion-dollar check imaging systems are not for every bank. The Mentis report said such systems are volume-driven and can only be cost-justified in a shop that handles 250,000 or more items daily. "A shop with less than 20 [check] encoders is just not going to see the benefits," Mr. Moore said.

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.