Text Size:

DT, January 2017

Next Up: Cognitive Payments
January 1, 2017

By George Warfel

From the early days of artificial intelligence in banking, back in the 1980s, we have come to where today’s cognitive technology is going to have a transforming impact on payments. Major changes in technology always create opportunities to rethink operations and strategy. The more radically different the technology, the greater the scope for change.

Cognitive technology creates the opportunity to develop a new approach to payments built on the technology’s human-like capabilities, and then to train a cognitive computer to implement the strategy and come up with its own additional changes as it gains experience. Cognitive technology learns through self-education. Its primary value is its accrued expertise. Both its expertise and its reasoning continue to evolve as new information is encountered and as it observes responses to its actions.

The changes that cognitive technology could enable at a bank will not always be obvious and may be different for each institution using it. Changes in payments strategy will be required to exploit the potential of cognitive payments. It is important to understand that the value of a cognitive system is not simply in replacing human labor. The value lies in its ability to conduct analysis and make decisions as well as or better than most humans would.

Let’s look at a couple of prototype examples:

A customer wants accelerated availability on a large inbound foreign payment. The cognitive system looks at the last 18 months of payments the customer has received and finds disputed/slow-pay transactions increasing. It checks if the payee has other accounts or assets with the bank that could back-stop funding of the payment, checks current regulations, and places the appropriate holds on the accounts.

The system routes its decisions to the customer’s bank calling officer with a summary of the relevant information. It then monitors the account, and when it sees the payment is successfully completed removes the holds on the other accounts and updates the account officer. The system may decide to recommend putting the customer on a watch list due to the pattern seen over the months. Finally, it stores the record of the incident and the actions taken along with the supporting documents.

Another example:

The cognitive system sees a sudden drop in activity in an important customer’s account. The system then discovers several things that might be connected with the drop in account activity, including a recent local news story about a person with the same name as the customer starting a new business, and an ad showing that a competing bank has a branch near the location of the new business. It reads the competitor bank’s recent advertising and sees that the competitor is offering free personal accounts for people opening new business accounts.

The system alerts the customer’s personal banker and shows her the pattern of activity in the account, the news clipping, the competitor’s ad, and a photo of the new business’s owner from a recent Chamber of Commerce publication. The system reminds her that the bank turned the customer down for a second mortgage two years ago, before the banker was assigned the account. In a matter of seconds, the banker has been alerted to the concern, has been provided with relevant information from multiple sources that help explain what’s going on, and provided a photo so she can confirm if the person with the same name as the person opening the new business is the bank’s customer.

Couldn’t the bank’s systems and employees be trained to notice these events and solve these issues? An employee couldn’t review 18 months of transactions in seconds. Current technology might not know what regulations and policies to check before placing the account holds. The bank’s current technology wouldn’t check for rival banks’ branch locations near the new business, and then check the competitor’s recent advertising. Would the average employee think to match the new business address to the Chamber of Commerce article and use the photo to confirm the owner’s identity?

Most current systems wouldn’t detect the alerting events, and if one did it probably couldn’t find the data, both inside and outside the bank, and determine what actions to take. Cognitive technology is designed to be trained to do these things, and to continue improving its capability.

Cognitive technology has been a long time in the making. Now that it is here, it will revolutionize payments.

George Warfel • GWarfel@haddonhillgroup.com

Share |


Read Digital Transactions Online
read more