- Key Insight: Citizens is using artificial intelligence to increase efficiency in nearly every aspect of its business.
- Supporting Data: The project, called "Reimagine the Bank," is expected to take three years and cost $300 million.
- Expert Quote: "The technology is in a spot at the moment where it's not hopes and dreams," said Citizens president Brendan Coughlin. "It's here and present right now, and we're leveraging it."
Brendan Coughlin became president of
That may sound ethereal, but the project the Providence, Rhode Island-based lender calls "
"The technology is in a spot at the moment where it's not hopes and dreams," Coughlin told American Banker in a lengthy interview. "It's here and present right now, and we're leveraging it."
The $226 billion-asset bank
"What we're trying to do is turbocharge the productivity of our people so that they can focus on the more challenging tasks and more important interactions with customers," Van Saun told American Banker last year.
Almost exactly one year into his time as president, Coughlin sat down with American Banker to discuss how
The following is a transcript of our conversation, edited for length and clarity.
American Banker: I want to ask you about the Reimagine the Bank project. So how's that project going? What progress has been made so far?
Brendan Coughlin: So, Reimagine the Bank I would describe as: 50% of the program is internally focused on positioning the bank for the future, on non-tech dimensions, and 50% of it is really leaning into AI and using modern technology.
We're one of the first banks to be fully on the cloud and now leveraging AI to reimagine our business model, front to back, in a really bold way. We're going to invest about $300 million in incremental technology capital over three years, with the goal of having an exit run rate impact of $450 million in earnings value creation by Q4 2028.
So it's going quite well. We're about six months into the effort, and we've sized up 47 unique initiatives that we're charging at across the bank, and I'm happy to share a few of those with you.
AB: Yeah, absolutely. So 47 is a lot to get through, but maybe you could tell me a few of the big ones.
Coughlin: On the AI front, I'll give you two examples. One is in our retail call center. We're going to deploy — we've actually started deploying; it's in pilot — agentic AI into our call center with the goal of 50%-plus of our phone calls will ultimately be answered by a non-human in a very effective way. And we expect to be halfway through that journey by the end of 2026.
So it's very real. It's in play right now.
AB: Could you tell me a little bit more about that? Have you tried this out yourself? What's it like talking to an AI bot over the phone?
Coughlin: The technology has advanced in such a way that makes this an incredibly positive customer experience. And so the voice technology has been tested in such a way that I would stop short of saying, "You can't tell at all that it's a non-human." But it is close enough that it's a very engaging experience.
And when you use an LLM sitting on top of it, you think back to how automation in call centers happened five, seven years ago. You had to say the perfect thing, and then the perfect thing led you to the right work stream. And it got clunky, and you wanted to hit zero and just get to an agent.
The introduction of an LLM that can actually understand what you're saying and be smart about how to route you has advanced in such a way that it's incredibly efficient and effective for customers to use it and get to the right place quickly. And then you can actually have that automation go all the way through.
So if you want to reorder a debit card, not only can it hear what you're saying, it can actually then connect to our back-end operations and actually order that debit card for you and send it to you.
And so it's a very quick call, a very engaging one that can then talk back to you in normal language as well, so it's not as robotic. So it's quite compelling.
AB: Could you give me a few other examples, besides ordering a debit card, of the kind of questions that customers could ask?
Coughlin: Everything from basic stuff — you'd be surprised how many people, in today's day and age, are calling to check their balance. You could just look at it on mobile, but a lot of people like to use the phone.
You could start the process to file a fraud claim, if there's fraud on your account, and get information back. You could change your address. You could call up and say, "What's my last 10 transactions on my account?" And that AI agent will look into your transaction history, be able to understand that and spit it back to you.
So we're building all these work streams. And of course, we said 50% will be answered by a non-human. So 50% won't, at least in the near term. So when you have something really complicated, you want to talk about retirement, or you want to talk about getting a mortgage or something more sophisticated, that's when our people will pick up from there and jump in.
AB: And I know that Bruce [Van Saun] has mentioned before that another area where AI could be applied is fraud prevention, anti-money laundering. Can you tell me a little bit about how AI is being used there?
Coughlin: Yeah, we're working on that now.
So if we're evaluating a risk profile for a customer coming in through [anti-money laundering], sometimes there's a lot of paperwork, sometimes you want to scan public-record information to see what's out there on a client. Using AI to go do that and come back with a summary report can give you a risk profile of public information. These are the types of things that we're looking at that used to be very human-intensive and research-oriented that you can now leverage AI to help you get to answers faster.
So it should be an effective way to speed up our processes, which is good for the bank, but it's also good for our clients. We can move much, much faster for our customers.
AB: Any other examples of AI in action?
Coughlin: I'll give you one more. Another human productivity boost would be an initiative we're calling AI-Powered Banker. And it's starting in our commercial bank, but it will quickly scale to other job families.
And that's an initiative where we're using AI to go get tons of data and information on our customers, whether it's internal data or external data, and package it up for our sales folks so they've got basically a health report on our client. …
So you can imagine a world where you even apply it to retail banking, where a customer's coming, you're walking into the branch, and somebody types in, "Tell me everything I need to know about Nathan." And up pops a one-page report: "Nathan gets paid on Friday. Here's his cash flow. It looks like his mortgage rate's too high. We should talk about a refinance."
An AI assistant can go in and find all that, and streamline it and simplify it and give recommendations back to our humans.
AB: So what's next? What's coming up over the next year or so for reimagining the bank?
Coughlin: No. 1, we're in execution mode. To your point, 47 initiatives is a lot, so the next year is going to all be about bringing this to life, moving from pilots to scale impact.
But we're not locking the program in at 47 initiatives. The capabilities are moving so fast, so we fully expect to add in new and bigger ideas as capabilities mature over time.
And so Job No. 1 is to make sure what we've bitten off we can actually chew and deliver. But Job No. 2 is to keep the periscope up and make sure that we're looking at the market and the world and how fast things are changing. So it's going to be an ever-evolving initiative over the next two to three years.
