From BAI’s TransPay 2003 conference
Image-Enabled Fraud Control

Imaging cost bank fraud experts the ability to smell or feel a bad check, but it gave them new tools as well as the opportunity to extend the fight back down the check-processing chain and ultimately to the point of sale, says Jodi Pratt, senior vice president, Risk Solutions, at Carreker Corp.

The tools may be having an effect. In 2001, survey data showed that check and deposit fraud rose by only 3%, after a decade of double-digit increases, Pratt said. Fraud rose from $679 million in 1999 to $698 million in 2001. In the previous survey, the increase had been 33%.

Unfortunately, while big banks seemed to be cutting check-fraud losses, fraud operators appeared to have shifted their target to smaller financial services companies. “The mid-size and smaller financial institutions saw tremendous upsurges,” she said.

“We’re raising the bar for fraud operators attacking the larger financial institutions, who are very often better positioned financially and technologically to attack fraud control, and we’re pushing it down to smaller organizations that don’t have the resources and are more vulnerable to that,” she said.

Further, Pratt said, the numbers suggest the size of the threat actually is increasing. In 1999, banks participating in the survey estimated that for every dollar lost to fraud, fraud control avoided $2.24 in losses. In the latest study, that figured jumped to $5.21.

“It’s telling us that the exposure - the number of attempts and the actual total dollar exposure -- is significantly increasing,” she said. In 1999, total exposure was $2.2 billion. In 2001, it was $3.6 billion.

“We’ve got to ask ourselves the question: Is the problem really going away, or is it just moving?” she said.

Pratt said that change itself works in favor of fraud operators. She said one of the lessons of the last decade is that the more banks add products or delivery channels or change payment mechanisms, the more they increase their vulnerability to fraud.

“When we’re introducing new technology and banking options, the people who are going to figure out how to use it most effectively first are the fraud operators, who have nothing to do with their time but spend 100% of it concentrating on how to break the system,” she said.

Pratt said that when bank fraud staffs see the total exposure number rise to a figure such as $3.6 billion, they fear that fraud operators will break through new delivery channels and payment mechanisms and start taking larger amounts of money again.

In the early 1990s, banks had started putting good fraud detection in the paper check-processing environment, Pratt said. “We could see frauds; we could taste frauds; we could smell frauds a mile away,” she said.

At the time, Pratt said, she was thoroughly alarmed at the prospect of imaging’s replacing paper checks. Today, she is one of imaging’s champions.

For example, in traditional on-us check-fraud review, suspect items are sorted in account-number order. That’s the way they come off the system, and that’s they way they need to go back. Unfortunately, fraud may occur on a big-dollar item far down the list.

“Maybe by the time the analyst gets to it - if they get to it - it’s too late to return the item,” she said. “It’s past the deadline, or you have to have a lot of staff members constantly available to do that kind of review.”

Image can change that. The image is available when the analyst needs it, and images don’t need to be kept in a particular order. Suspect data can be re-prioritized based on dollar exposure so that those items are examined early in the morning.

Imaging can bring a huge productivity improvement, she said, and often that is translated into staff reduction. But she said that in fraud control, it might make more dollars and sense to retain staff.

“What you might find is that you can save 10 or 20 times the cost of the salary if you keep the person and allow them to do more of the review on fraud control,” Pratt said.

Image has another advantage in on-us review, she said. Typically, when analysts pull a suspect item, they want to pull a companion item as well so they can compare handwriting or other patterns. One problem is that there often is no companion item. Another is that fraud operators have learned how to create a companion item that will pass through the system.

In an image environment, the analyst has access to companion items from a month ago or other periods during which the customer has reviewed the account and, in effect, confirmed that the check or other suspect item is good. Similarly, on the deposit side, the analyst can see not only the deposit slip but the transit items, including the signatures and data on them.

The advantages of images go beyond improving current processes, Pratt said. Images also offer new detection opportunities.

Up to the early ‘90s, someone who wanted to pass a counterfeit check pretty much needed to obtain a real check, copy it, change it and try to make it look legitimate, she said. The copies usually were of poor quality, accounting for bank fraud staffs’ ability to spot them relatively easily.

Today, desktop publishing enables customers to print their own checks. Magnetic ink is available at Staples. Unfortunately, some legitimate customers decline to pay for MICR ink, and their checks are rejected like fraudulent items. Or they produce checks without a standard format, so the system kicks out a lot of false positives. Meanwhile, fraud operators who know their business can create a good-looking check that a banker no longer can feel.

Luckily, image offers some help. First, machines can take over much of the pay file process in which a staff member compared the signature on a check to a signature on a card on file at the branch. Signatures from the cards can be digitized and used to compare against the digitized signatures on checks being processed. The ability to process hundreds of checks per second offers banks the option of comparing signatures on low-dollar amount items or even all items.

Another innovation is check stock recognition, a process that compares non-written information printed on the check stock itself. Comparisons might be made on the size of the check, its borders, or the font used, Pratt said.

Image also enables a bank to move some fraud control from Day 2 (back office) operations to Day Zero, or close to real time, at the teller window or ATM. Capturing the image at the point of presentment enables the bank to bounce the image against all its internal fraud controls - positive pay, signature verification, check stock recognition, etc. Before the money goes out the door, the teller can get a message not to accept the check, or a supervisor can put a hold on the money.

Imaging also removes the need to have an experienced supervisor at each branch to examine checks over the review threshold. Images can be routed to a supporting site or a centralized review area.

Imaging also can improve how banks process losses. It offers the flexibility to simultaneously:

• Retain account-opening or other data for long periods, as may be required by the USA Patriot Act
  
• Maintain all associated case data together
  
• Keep originals safely stored
  
• Keep the chain of custody of originals intact
  
• Enable bank staff or law enforcement personnel access to all associated data as often and as long as needed
  

Banks also could move internal fraud control out to other banks, for example by using images to verify each other’s official checks. Ultimately, verification could be pushed out to retailers. A check image could be verified through the fraud filters at the bank on which the check is drawn “before the stereo goes out the door,” Pratt said.

Pratt said she didn’t think many banks have approved an image project primarily justified on risk management. But she said risk reduction can be quantified and can “beef up” an image business case.

For example, if a customer pulls $100 out of the bank, the bank loses a few dollars of opportunity cost, she said. But if the bank loses $100 to fraud, every dollar of it falls to the bottom line. “Anything you can do to protect against fraud loss is a direct benefit to your business case and to the bottom line of your company,” she said.

Another approach is to go beyond imaging’s transaction benefits to bolster the business case. 9-11 gave federal agencies support to call for additional bank security. “Maintaining identification on everybody that you begin a relationship with would be a really onerous task in today’s environment without image,” she said.

New revenue opportunities always boost a business case, Pratt observed. She mentioned “reverse positive pay.” Some small and mid-sized businesses can’t commit to creating an electronic file and sending it to the bank each day as required by positive pay. But if the bank sent them a file of images cleared the night before, some of those businesses might jump at the chance to have a bookkeeper look for suspicious items.

Bankers who are trying to make the case for image might consider these next steps, Pratt said:

• If you’re from the image side of the bank, have a cup of coffee with your risk counterpart, or vice versa. Identify where you’d like to go. Look for ways to work together to get there. Ask: Where is the bank going with ECP and Check 21? What kind of national fraud databases does the bank participate in? Is the bank considering protection services for merchants?
  
• Develop a migration plan. “Neither a risk platform nor an image platform - and certainly not the two together - is a short-term project,” Pratt said. Identify everything you want to accomplish, but focus on what you want to do in the next 12-24 months. Identify expected benefits that will accrue along the way.
  
• Keep relevant technologies in mind as you discuss image-enabled fraud control. Does the bank have an intranet? How are you going to move images among units of the bank? Are you going to use public channels to move information? Do you intend to put scanners on the front line? Do you want to do POS fraud verification? Do you want the tellers to do real-time verification? Does the bank have the infrastructure to support this? Is a needed upgrade on the drawing board?
  
  

Clint Swift
BAI’s TransPay 2003 conference was held May 7-9, 2003, in New Orleans.