The world is not soon going to be taken over by robots — and neither are credit unions. But for the latter, bots can be a crucial component of member engagement and employee productivity.
So argued Matt Kinney, founder and CEO of Attando Technologies, a provider of artificial intelligence solutions to credit unions, speaking Sept. 12 at the Credit Union National Association (CUNA) Technology Council in Chicago.
“You can’t go to a conference in 2019, without hearing the hype associated with AI,” said Kinney. “But I want to make sure we don’t overpromise and underdeliver, because there’s a lot of that going on.”
Instead, he suggested that credit unions focus their AI development on four distinct use cases:
1. Chatbots Are Central to Credit Unions’ AI Strategies
Chatbots are the among most common applications of AI inside retail banks and credit unions, and for good reason: Customers have come to expect a convenient, frictionless, conversational experience with their financial institutions at all times. They’re interacting daily with Amazon and other digital-native businesses, and those interactions have reset their expectations.
Kinney noted that in his home state of North Carolina, he was able to use a text-based application recently to renew his automobile registration with the state’s Department of Motor Vehicles.
“They made it really simple, and — dare I say — an even pleasurable DMV experience,” he said. “My point is, even at what we think of as the lowest common denominator, this is the kind of experience your members are expecting to have.”
To make their deployments more successful, Kinney said, credit unions should set customer expectations fairly. A chatbot on a credit union website is not Amazon’s Alexa: It can’t tell jokes or answer history questions. When done right, it will be a specific-use bot, well trained to manage a customer or prospect through common issues.
Still, he said, a good chatbot should be able to go beyond the basics. It should be able to authenticate members and potentially help them with their own accounts; to do that, the AI needs to be integrated with the institution’s core banking system. It should also be able to turn the interaction over to a human agent when necessary.
“The value is really in the authenticated experiences,” Kinney said. “It’s great to be able to do FAQs, but the greatest value is being able to answer real members’ questions, to be able to authenticate them as members and handle their real problems.”
2. AI-Driven Co-Pilots Can Save Credit Union Members Time
Like any financial institution, processes within credit unions can be cumbersome and difficult for members to navigate.
An AI-driven “co-pilot” can help members through them. Such a use case is helpful when a credit union has identified a specific problem that artificial intelligence can help solve. For example, Kinney said, an online lenders’ members were abandoning its debt-relief application before completing it, so they built a co-pilot — essentially, a special-purpose chatbot — to intervene when prospects seem to be wavering. The bot is trained to proactively provide information that would help the customer continue with the application, and to answer relevant questions.
3. Agent-Assist Bots Help Credit Union Employees Provide Better Service
When a customer or prospect interacts with a live agent in a call center or chat room, an agent assist bot can monitor the interaction and offer the agent appropriate canned responses or other resources to help the agent successfully navigate the call.
4. RPA Can Increase Credit Union Employees’ Productivity
Robotic process automation is the use of bots to complete repetitive and typically laborious tasks, enabling humans to focus on other things. RPA is growing quickly in popularity, especially among financial institutions where employees perform a large amount of routine data entry.
But there are some pitfalls with RPA, says Kinney, including the tendency of credit unions to sometimes “chase shiny objects.” Because RPA is hot, executives sometimes want to deploy it before clearly identifying how it should fit into their overall workflow. “The CEO says, ‘We have to be doing this,’ and they struggle because the next week, it’s blockchain,” Kinney said.