 | ||||||||
| ||||||||
| | ||||||||
| In partnership with
Hi! Kiah here. Welcome to Fintech Takes Banking, my weekly newsletter where I highlight things I think are interesting or important for bankers and the surrounding environs. Was this email forwarded to you? Sponsored by Creatio How is AI changing financial institutions’ operations? Financial institutions have moved past AI experimentation and are using agents to expand workforce capacity, augment teams or create new roles, according to the second Financial Services Edition of its State of Enterprise AI Adoption report from Creatio. Creatio surveyed 190+ financial services IT and business leaders to see how their institutions are scaling AI and AI agents. They found three major shifts shaping the future of AI in financial services: 1. AI is reshaping the workforce — not reducing headcount. 2. The value of AI is expanding from productivity to revenue growth. 3. AI adoption leaders are setting new benchmarks for enterprise AI. Check out the full report to see where AI is having the greatest impact in financial institutions worldwide. Why 3 Banks Took on the Dreaded Core ConversionThe story of modern banking is often, surprisingly, a story about cores. I won’t belabor the role a core plays in a bank too much; you probably work at a bank and live and breathe this in a way I don’t. But one truism in banking is that the core is often a bank’s most important, most expensive and longest-duration technology — and many banks don't like their cores. In a 2024 American Bankers Association survey on cores and core providers, just over half of respondents reported being extremely or somewhat satisfied with their cores. One reason for the lack of satisfaction could be the age of the core, with 68% of 2024 respondents saying their bank had been on its core for 10 years or more. Bankers’ satisfaction with their core declined the longer their institution was on the platform. More than three-quarters of respondents at banks with more than six years left in their core contracts were somewhat or extremely satisfied, compared with 41% of those with two years or less left. That puts these unhappy banks, with their potentially outdated systems, in a tough spot. Bank customers increasingly conduct business online and want real-time functionality and personalization. Newer companies can use modern technology, powerful data analytics and now artificial intelligence to offer deposit-like products and payment services to woo bank customers away. Banks want to offer modern products and services, but their cores increasingly keep them from competing. Very few banks decide to change their cores in response to these pressures, or to a deteriorating relationship with their core provider. I talked to three. The three banks are very different. Republic Bank & Trust Co. in Louisville, Kentucky, has been live on its new core for a little over a year. Flagstar Bank in Hicksville, New York, and Emprise Bank in Wichita, Kansas, picked new cores this year and are in the middle of their conversions. I spoke to executives at each institution to learn about their journeys and any advice they have for bankers at other institutions facing similar choices and conundrums. Why’d You Do It?Each bank had its own reason for electing to convert, but executives at all three felt their cores weren’t built for the future. They wanted real-time visibility and capabilities, a better core partner and a platform flexible enough to let them innovate and respond to change. Republic’s core migration journey started outside the core, with its cards, said Jeff Starke, the bank’s chief information officer and chief operating officer. He said the bank was “not having a quality experience” with its card solution and provider and ultimately decided to issue a request for proposal, or RFP, for a new card program. Executives knew that searching for a new card partner could scope out to include the core, so even though they didn’t set out to replace the core, they included it in the RFP. Republic, which is the bank unit of the $7 billion Republic Bancorp, invited the top three contenders to participate in a demo day, with about 100 associates from around the company assessing each product and system on a scorecard. Those results surprised the executives leading the process: employees wanted to move to a new core. “Bankers, maybe like most people, don't love change, and they're not likely to suggest that they want to go through a core conversion,” Jeff said. “So that was a little surprising to see, but that was the result we were left with. Couple that with the service interruptions and the lack of the technologies and services we felt we needed for our clients and that drove our decision to make the change.” Flagstar Bank’s core conversion is part of a broader transformation project, said Chief Technology Officer Jason Pope. Flagstar is the product of three legacy institutions — New York Community Bancorp, Flagstar and Signature Bank — and was running two cores. In early 2025, management at the $87.4 billion bank wrote a 174-page document outlining its vision for its “simple and sophisticated,” or S2, platform, and the work needed to build it. First up was consolidating six legacy data centers into two newly built hyperscaled and fully mirrored data centers, which took 11 months. The bank also inventoried its two cores — every application, interface, batch job and report that touched the core platform — and decoupled them and the underlying business logic from the core. Jason said this added up to “hundreds” of integration points, but it would let the new core do only what a core is meant to do: hold the record and process transactions on the ledger. Before starting its RFP process, the bank also adopted a product operating model, mapping the products and services Flagstar offers and the experience it wants clients to have. That informed the technology requirements it included in the RFP. “We started by writing down the platform business capabilities. Not a feature checklist: real-time postings, product configuration without code, clean API access to the data and our ability to scale the bank as it grows were fundamental,” Jason said. The legacy core at Emprise Bank, the bank unit of Emprise Financial Corp., had been showing its age for a decade: it uses batch architecture, it’s not cloud native, it doesn’t have robust APIs and it’s account-centric, not customer-centric, said Chairman and CEO Matt Michaelis. The cost of those constraints has grown over the past 12 months as executives thought about customers' changing expectations and how the $2.9 billion bank would be able to deliver personalized experiences in real time. Matt said Emprise sees a new core as a way to preserve its “strategic freedom and flexibility,” helping it stay agile and navigate a rapidly changing world. He has spoken with many bankers who are interested in wrapping their legacy core in an agentic AI layer, and that might be the right answer for them, he said. Emprise ultimately decided that while a wrapper would be beneficial, it wouldn’t fix the batch processing or integration challenges that currently exist. What Did You Pick, and Why? Flagstar’s RFP process took several quarters and involved almost a dozen candidates before selecting the Finxact core from Fiserv, which it announced in August. Finxact is a cloud-native headless core that offers financial institutions modern API architecture, said Srini Krish, cohead of financial solutions at Fiserv, which also provides Republic’s core. It connects and integrates with any type of payment system or servicing platform. Jason added that real-time settlement and processing give the bank better fraud detection and richer data for AI-powered tools. Matt said Emprise conducted a relatively thorough search for a modern core that offered full solutions for the bank’s business units, but these requirements narrowed the list of potential partners significantly. Emprise felt strongly that it needed a modern core with AI workflows incorporated or built on top of it. Ultimately, he said it was “a fairly easy” choice to select AI-native core banking technology from Maximum, which it announced in October. Maximum, which the bank had previously worked with, is a banking infrastructure company founded by former Marqeta executive Randy Fernando that went live earlier this year. Matt said the bank was comfortable with how the company builds its technology and integrates with the bank. He also said the migration will happen at “a reasonable pace.” Emprise expects to run a “very small,” contained use case on live traffic that it will shadow test by the end of the year before moving more production over the course of 2027. The new core will allow Emprise to scale to 10 times as many customers in the next 24 months and improve its efficiency, according to an October article in FinTech Futures. “Any banker that's been through a core conversion knows how painful that can be, and I think there's a lot of hesitation in the industry to engage in a core conversion because it is so difficult and there's customer impact,” Matt said. “I also believe that as the technology tools have evolved and are continuing to evolve, we can make it somewhat less painful… I'm optimistic that it won't be easy, but that it will have less impact and be smoother than what we have experienced in the [past].” Republic Bank selected Fiserv’s DNA core and has been live for a little over a year. Chief Transformation Officer Jenna Cowley, whose team came up with and executed the conversion, emphasized that Fiserv has been a strong partner throughout the process and brought in various experts to work on site alongside the bank. The bank also engaged with the Smith Consulting Group, which is now a part of Fiserv, to assist on the conversion. Jeff said that Republic management made it clear that the core conversion was a company initiative — not an IT or operations project — and everyone needed to be on board. The alignment, and the work of Jenna’s team, paid off. Republic went from “contract signing to live in eight months,” said Jeff. I thought that was fast, although Fiserv’s Srini noted that one bank converted its core in four months. The difference in the cards program is “night and day,” Jeff said. The bank went from needing to address issues on the debit or credit side — sometimes both — daily, impacting operations and distracting management, to a very stable card environment. Outside of cards, the core has a modern architecture and robust API structure that allows the bank’s internal development team to innovate. And Jenna said her team appreciated the opportunity to “clean house” and reevaluate outdated processes and procedures. What Advice Do You Have for Others?Most banks, when given the option, don’t plan to convert their core — even if they are dissatisfied. Only 19% of respondents to the ABA survey’s said it was somewhat or extremely unlikely that their bank would stay with its core at the next contract renewal, compared with 69% who said it was somewhat or extremely likely they would renew. Until now, many bankers seem to ascribe to a "devil you know" approach and that the trouble of a core conversion isn’t worth the benefits. That may change as technology widens the gap between what’s possible and what a bank’s core permits. Maybe more banks will decide that after 10 years, they need more modern infrastructure and architecture so they can take advantage of more powerful data analytics and agentic AI. “Obviously there is a ton of risk in a core conversion,” Jeff said, adding this was the 18th core conversion in his banking career. “It's one of the primary reasons that banks don't tend to do it, even if their current core isn't giving them everything they want or their current relationship isn't giving them everything they want or need.” But staying with outdated technology carries risks, he pointed out. Many banks right now use strategies to work around their cores, which adds extra tech, work and complexity. He encouraged executives to be realistic about evaluating their cores and “unafraid to make the change when it becomes obvious.” He also recommended that executives factor their relationship with their current core provider into any assessment. Some relationships have become dysfunctional, and it may not be worth trying to salvage them. Matt at Emprise advised bankers to start by being “very clear about what you want and need. What are you solving for? What do you hope to be able to do?” He suggested stating these as business needs, rather than tech requirements, which will help executives identify what they want their core to do and evaluate what and how much should change. He also encouraged executives to be future-minded: Of course, they need to serve customers well today, but they should also think deeply about what the institution will need to differentiate itself and remain relevant as it moves forward. Jason at Flagstar urged executives to treat a core conversion not as a technical program but as a “people program,” since it involves “everything” in a bank’s ecosystem. This approach emphasizes the role of change management and communication that can help these programs run more smoothly — or derail them. “Core conversions most often fail not because of a technical or an architectural problem; they have challenges at the change management layer,” he said. He also urged management teams to take the long view and avoid treating a core conversion or consolidation as “a one-time event.” Instead, executives should think about what the next three to five years could look like and what they may want from their core. Going Through ItNone of the executives I interviewed would tell another bank to convert its core just to fix its technology problems. They all acknowledged that converting a core is disruptive, expensive and carries risk. Still, they understand how many banks on legacy technology feel right now: that their core constrains them, and those constraints are leaving them behind. I normally file core conversions in the same category as root canals and cross-country moves: involved, expensive, tedious and sometimes painful. And while none of the executives would say the experience was easy — in fact, two are still going through it — I couldn’t help but notice how excited everyone was to tell me about the conversion and their bank’s new core. Usually, I’m the energetic one in an interview, and my sources are more tempered. But I could hear the enthusiasm and pride as I spoke with Jason, Jeff, Jenna and Matt. They’re bullish on their institutions’ futures and the technology that will power them. And that’s something you rarely hear from bankers when they’re discussing their core. Sponsored by ElevenLabs Banks need AI agents focused on how to best resolve customers’ requests, get the details right and hand off to a human agent when they should. FROM THE VAULT What’s on my mind and filling my time: 🎰 Do celebrities endorsing prediction markets not have enough money already? Amanda Mull in Bloomberg asks the question I always find myself wondering. Also I don’t keep up with the NBA but it’s so cool Wemby refuses to do these endorsements (and his Halloween costume rocked). 🎤 Bonus Podcast Rec: How should banks respond to customers’ AI agents? In episode 8 of Agent of Record, produced in partnership with Parlay Finance, 37Maru’s Frank Rotman joins me and Alex McLeod to discuss how agentic AI could completely disrupt the consumer inertia that underpins bank profits. We explore the problems that AI agents address for consumers and the problems they create (liability!) how agentic search changes how banks will be discovered and whether institutions should respond by creating their own agents. Listen now!* 🏥 Public health is so cool! The New York Times published an article about an 80-cent intervention that helps medical personnel in developing nations identify whether a woman who has just given birth is experiencing postpartum hemorrhaging. (Warning: there is a mild photo depicting blood.) 🎙️ On Closing Credits: Catherine and I are joined by Brian Fritzsche, associate general counsel at the Consumer Bankers Association, to discuss the 2002 Stephen Spielberg film, “Catch Me if You Can.” 🛫 Catch me at: AFC Policy Summit on Nov. 17 in Washington. Piper Sandler’s Balance Sheet Strategy Seminar on Dec. 7 in New York. *This rec is brought to you by one of our fantastic brand partners. Thanks for reading! Let me know your thoughts. - Kiah | ||||||||
|
