A Regions Bank executive took advantage of the visit of a group of college students to do a little ad hoc research. He asked, “What frustrates you the most about dealing with the bank?” One answered: “My cell phone knows more about me than you do.”
That was several years ago, says Michele Elrod, EVP and Head of Marketing for Regions, and two things have changed since then: Mobile phones know even more about us than they did then and Regions Bank has greatly improved its customer knowledge. The southeastern regional bank invested substantial money and talent to build two artificial intelligence platforms aimed at improving customer experience, engagement and revenue.
One of the two platforms, nicknamed “Rosie,” achieved an internal rate of return (IRR) of nearly 50% for the first phase of its implementation, according to Elrod. This compares with an IRR for most tech projects at the bank of between 12-15%.
The second AI platform, built around IBM’s Watson software (and nicknamed “Reggie,” internally), paid for itself in the first year by a nearly half-minute decrease in hold times at the bank’s contact center, according to Chris Brasher, Regions’ Head of Bank Operations, speaking at an IBM event.
Although the two AI platforms were developed on separate tracks, they will increasingly work together, according to Elrod.
Why Marketing Has to Embrace AI Software
Regions Bank, the main unit of $129 billion-assets Regions Financial, didn’t set out to incorporate artificial intelligence into its marketing, which is how things turned out. According to Elrod, who spoke at The Financial Brand Forum, the initial goal was simply to get smarter, faster and more efficient in marketing and improve customer experience.
Elrod, who has been in bank marketing for 35 years, realized several years ago that “we could not continue to talk to the customer the same way that we always had. We could not market at them. We had to market with them.” Marketing now really is about having a conversation with the customer.
Elrod believes traditional marketing still has a role. “It’s important to manage your brand at a holistic level,” she says, “but one-on-one marketing and personalization are key now.”
As she told The Financial Brand in an interview, “People don’t wake up thinking about banking. They think about money, they think about their life and what they need to be able to get where they want to go. I firmly believe we should always look at what solution we are bringing to consumers and speak in their terms.”
“That doesn’t mean you abandon marketing your products and services,” Elrod continues, “but you put that in the perspective of solving a need. You develop relevant content. You talk with people and bring them along.”
The same technology driving changes in consumer behavior has created huge new opportunities — and challenges — for Marketing. “We knew that we had to change everything about the way we did marketing if we were going to compete for consumers’ attention,” states Elrod.
That required building a data management platform and analytics capabilities, and adding new skills. In the past year the bank has hired a data scientist along with other specialized talent, including a channel strategist and a customer interaction specialist. All told, there are four people dedicated to the Rosie project within Marketing, plus two other dedicated staffers in Marketing Operations.
- How Two Banks Are Chasing the AI Dream
- 5 Tips For Financial Marketers to Tap AI’s Personalization Potential
Rosie Instantly Provides ‘Next Best Actions’
In her Forum 2019 presentation, Elrod spoke primarily about Rosie, which was developed by her marketing group. Reggie, the Watson-based platform, was developed elsewhere in the bank and is used in the contact center to help route people faster, and solve problems faster.
Rosie is actually an acronym, standing for “Regions Optimal Solutions Intelligence Engine.” It (or “she”) is a centralized decisioning engine comprising 22 models that ingest data, insights, analytics, content and more from a variety of marketing platforms and databases. Regions built the engine using IBM’s Interact software — i.e., Rosie does not use Watson technology.
Rosie gives the bank’s branch and contact center staff the ability to speak with consumers on in-bound communications in a more relevant and timely way, prompting them with one or more “next best actions.”
Such “NBAs” are the key deliverable at this stage of Rosie’s development. These prompts show up within milliseconds after the initial customer contact is made. In general terms NBAs answer the question: “What should we talk to this customer about?”
Next best actions “are not about hawking products,” Elrod states. They’re about building customer engagement and loyalty. She says 40% of NBAs are not related to selling. These are informational. Things like “Do you know you have 5,000 rewards points that will soon expire?” Or, when a customer is cashing a large check at a branch, the teller would see an NBA suggesting loading the funds onto a card, which would be more secure. An NBA could even be an emergency preparedness reminder during hurricane season.
Elrod says NBAs don’t exclude products and services by any means, “but we have to include other useful information if we’re going to build trust and respect and engagement.”
As currently set up, Rosie produces three NBAs for a contact center person, and they choose. “We don’t dictate what they do,” says Elrod. “This is about giving them more information and allowing them to make more informed decisions about what to talk to the customer about.”
Education And Results Overcome Pushback
Regions’ marketing group maintained a control group to help measure Rosie’s impact. Elrod says that, on average, “we produce somewhere between 7% and 8% incremental revenue per customer if we touch them with a Rosie NBA.” Figures like that get the attention of the bank’s business groups, she says.
Yet Elrod doesn’t shy away from describing some of the internal discussions over Rosie as a “battle.”
“For the first year of Rosie , we had many, many talks with business units,” says Elrod. “They would say to us, ‘What products are in the next best actions? ‘How many times does each product appear?’ ‘What’s the first next-best-action?’ etc. You don’t want to start manipulating the models,” says the senior marketer. “The NBAs have to be relevant and timely for the individual customer, based on the inputs and the models.”
Two steps helped Elrod and her team negotiate the tricky waters of inter-bank cooperation. First, from the outset, they had the support of the head of the consumer bank.
Second, the bank formed a steering committee, which includes representatives from Operations and Technology in addition to Marketing and Analytics. It also includes Audit and Legal and Compliance. “If you say ‘AI’,” Elrod explains, “regulators’ ears perk up. They want to know, ‘How are you using that?'”
- Data is Powerful, But Not Banking’s ‘Silver Bullet’
- AI Could Destroy Traditional Banking As We Know It
Next Steps at Regions Bank for Artificial Intelligence
In the near future, a closer integration between Regions’ two AI platforms is slated, says Elrod. One key function will be to assess consumer sentiment or tone. As Elrod describes it, if a person is in touch with the contact center and Rosie is about to give the NBAs, Reggie will ping Rosie and identify “Negative tone. Anger tone,” and Rosie will instantly consider that and possibly change the NBAs as a result.
Rosie currently provides NBAs for all customer segments: Consumer, Commercial and Wealth and Elrod says Rosie eventually will be used in all channels — ATMs were just brought on line. (Customers are not aware of Rosie — the bank has not marketed the name.)
Regions will integrate Rosie with Salesforce to be able to start suggesting actions directly to bankers in the bank’s commercial call center, says Elrod. “Our vision for commercial is to not only provide the next best action, but useful information for that phone banker.” That might be an article about a particular industry that banker covers that is experiencing mergers. “It is important to note,” she adds, “that NBAs are solutions and solutions are not always products, they may be informational and educational.”
So far, Rosie has been applied to inbound communications. Elrod says she wants to apply the AI platform to work on inbound and outbound marketing together. “Why wouldn’t I want outbound decisions influenced by the same set of models and tools? Our ultimate goal is to have one view of the customer.”