EXCLUSIVE—  Artificial intelligence can produce the necessary insights banks need to provide the best experiences to their customers, but these insights depend, at the start, on what type of data the software is analyzing.

“Artificial intelligence is only as good as the data,” Adam Nanjee, senior vice president, digital banking, for strategic price provider Zafin told Bank Innovation. “Right now in the banking world, AI is so powerful because there is so much data.”

Zafin is utilizing AI at a “micro” level across most of its platforms, including Zafin Origin and Zafin Cloud, Nanjee said, an approach which allows the company to provide predictions to its banking clients at a more specific level: like determining what specific segment of a bank’s customers are in need of an auto loan, for instance.

The company currently works with 35 banks for this purpose, for both retail and commercial insights, including banks like Barclays and SunTrust, Nanjee said. This can be “really valuable,” he said, as some customers are “just looking for a checking account or a credit card” and not broader, less personalized offers.

Zafin’s platforms don’t aggregate any data of their own for these predictions, and don’t save any additional data, these predictions are fully dependent on the data sent over by the banks, Nanjee said.

As they rely on data sent via a single bank, does Zafin get a clear enough picture of a consumer’s financial life to make product predictions? Yes, Nanjee said, adding that the company’s machine learning technology can run down “to a specific region, a specific customer segment” for the bank to focus on when it comes to lending and consumer products.

Zafin added in a statement sent via email:

Typically, items like checking, savings, mortgage and credit cards all live on different systems within a single bank…Our software provides our customer (the bank) with an analysis of transactional data — spending patterns, for instance.

The customer’s banking activity with other financial institutions would not apply here — although conceivably, if their primary checking account is where they move money to/from, a bank could see how much money is leaving and to which other institutions. Since they rarely dive down to a single customer view — AI could be deployed to signal opportunities in this sense.

Want to continue the conversation? Join Bank Innovation’s Telegram here.