ChatBots: The delicate balance between Hal and ‘hello!’

ChatBots: The delicate balance between Hal and ‘hello!’

CUs should capitalize on their inherent advantage of personalized service.

April 12, 2017

One of the most promising and closely followed areas of fintech in recent months has been the application of artificial intelligence to deliver personalized financial services.

Such solutions, often referred to as ChatBots or robo advisors, offer a rare opportunity to transcend hyperbole and truly transform the provision of financial services.

Consider retirement planning and investment advice, areas to which such solutions are frequently connected.

It’s no secret that a disturbing percentage of Americans have insufficient savings to cover their post-earning years. This is likely the No. 1 area where financial institutions could step up to improve their customers’ financial lives.

Yet that very lack of accumulated balances is the complicating factor that keeps most financial institutions away: A shortage of assets means a shortage of fee opportunities, so these low-margin customers are shunted to self-service models.

And because these are the customers most likely to be intimidated by financial planning, they become frozen into inaction—and the cycle perpetuates.

Now imagine a financial discovery process being hand-held by an informal dialogue via text or voice, supported not by a live, on-the-clock employee but rather an intuitive interface able to gather and dispense personalized info in real time.

We’re not talking about filling out a static online survey instrument and spitting back a canned set of recommendations. Through machine learning, these ChatBots can return unscripted answers (i.e., not merely responding to magic keywords like “get balance”) and develop greater understanding of the customer’s preferences with ongoing engagement.

This concept need not be confined to big, weighty questions like retirement planning, either. Startups like Clinc have developed virtual assistants that plug into mobile banking to answer impromptu questions like, “Can I afford to buy a shirt for $150 right now?”

Better yet, some of these apps—those from UK firms Moneyhub and Meniga, for instance—leverage gamification and peer pressure to encourage savings and other desirable behaviors when it detects certain balance levels, unexpected outflows, etc.

Such robo advisors hold great potential to level the playing field—in both directions.

Through innovative start-ups, the technology will be available to credit unions and community banks at the same time as the big banks. And arguably, implementation may even be easier for the smaller players.

On the flipside, a well-executed ChatBot could help large banks narrow the high-touch service gap with their more community-minded brethren.

As with all such big new ideas, however, the path to Main St. isn’t nearly as straightforward as the initial headlines suggest.

For starters, credit unions need to be hyper-vigilant not to relinquish their well-earned reputation for high-touch member service. A robo solution will not appeal to everyone, and should not be pushed to all member segments.

Even for likely adopters, careful communications are in order to make clear this is not just another voice response unit. By the same token, painstaking advance testing should be conducted to ensure that responses (whether audio or text) do not seem stilted.

Another key is to build a process that stands ready to execute a handoff to a live representative.

Large banks may strive for end-to-end automation of many of these transactions. Credit unions should capitalize on their inherent advantage of personalized service, making a seamless transition to the live channel readily available at any point.

Think of leading ecommerce websites that monitor for periods of inactivity or apparent clicking confusion, providing a “Want to chat live?” offer. Navigating these transitions will be more art than science, and subject to their own learning processes.

In any case, it is essential that all data provided by the member during an automated session transfers to the live assistant. Forcing the member to retrace steps ensures a negative experience.

Finally, there’s perhaps the most vexing challenge: The entire premise of robo advice is to help customers make good decisions. The “best” decision for a customer and for the institution may not necessarily be aligned—think fee generation opportunities or the accumulation of revolving card balances.

Credit unions should be exceedingly careful in thinking through the recommendations it dispenses through robo advisors, perhaps presenting members with alternatives and summarizing their implications.

Of course, these same caveats apply to advice provided in live settings. However, recommendations delivered via computer algorithm carry an added layer of intrigue.

Mark my words: We will see a bank ensnared in some controversy on this front in the coming years. Credit unions will want to be on the right side of any such story as the “trusted partner.”

Advancements being made in robo advice are fascinating, and deserve your attention.

Credit unions will need to navigate a fine line, however, harnessing the benefits they can offer in advising members while minimizing the robotic aspects.

GLEN SARVADY is managing partner at 154 Advisors and senior payments expert with Best Innovation Group, a CUNA consulting partner. Follow him on Twitter via @154Advisors. His views do not necessarily reflect those of Credit Union National Association.