What It Takes To Scale A Lifestyle Banking Bot

business-ai-adoption

Lifestyle banking, including conversational banking components across web, mobile, messaging and voice channels, could be the wave of the future when it comes to consumer financial management. But first, it’s going to have to see widescale implementation and adoption.

That’s very different from rolling out lifestyle banking experiences to a few hundred customers in a pilot program. Some of financial services’ largest incumbents have invested heavily in artificial intelligence (AI) to power lifestyle banking innovations, only to find that full-scale rollout is taking more time and effort than expected.

Many of these players are now surrendering the reins to FinTech companies with expertise in wide-scale deployment. Kasisto is one such FinTech, with experience bringing mature conversational AI platforms to market on a scale that reaches hundreds of thousands, or even millions, of customers.

Kasisto CEO Zor Gorelov said that, for most banks, it makes more financial sense to license an AI from a FinTech than to build their own.

A bank’s IT team could build, test and deploy a bot that demos well, Gorelov said. But then, upon rollout, the bank could find that customers are asking different questions than the developers anticipated, or using different language than expected.

Data surrounding these differences must be collected, enriched and fed back into the assistant, said Gorelov. The bank must invest in not only the bot itself, but in the infrastructure, tools and staff to keep it running in tip-top form.

The problem there, said Gorelov, is that most IT teams are geared to build a solution and then move on to the next problem, whereas a bot is “like a living organism” that needs constant grooming and training.

“You don’t hire call center agents without training them,” Gorelov said by way of comparison. “Virtual assistants are no different.”

Yet it’s not just about hiring the best and brightest AI talent either. Even if a bank manages to attract 50 doctorate-level AI engineers, what happens to that team after the bot deploys? It’s a waste of good talent for doctorates to spend their time tagging code, said Gorelov — and it’s also not cheap for the bank.

He said that’s why Kasisto’s platform connects banks with qualified engineers to write the code, while also ensuring that someone in the organization knows how to oil the machine on an ongoing basis.

“You don’t need AI rock stars,” Gorelov concluded. “You can use your existing operating model, cost structure and talent.”