IBS Intelligence Global FinTech Interviews
Go one-on-one with the innovators, disruptors, leaders, and decision-makers driving change in FinTech and financial services. IBS Intelligence delivers exclusive global interviews that uncover strategies, challenges, and the ideas powering the next wave of financial technology.
IBS Intelligence Global FinTech Interviews
EP1023: Talking strategy, tactics and evolution
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
This interview features nCino CEO Sean Desmond discussing the company's strategic focus on integrating agentic AI into the global banking sector. Desmond introduces the concept of a "dual workforce," where specialized digital agents handle repetitive, data-heavy tasks to allow human bankers to prioritize client relationships and complex decision-making. The text highlights nCino's expansion into ambitious markets like the Middle East and Japan, while emphasizing that their technology is built on a decade of industry-specific data. By embedding AI directly into existing workflows rather than offering it as a separate tool, the company aims to significantly speed up commercial onboarding and enhance operational efficiency. Ultimately, the source illustrates how intelligent automation and unified data platforms can help financial institutions outperform modern customer expectations.
Imagine a commercial bank is um just days away from closing this massive multimillion dollar lending facility for a major corporate client.
SPEAKER_01Right. High stress, everybody's watching.
SPEAKER_00Exactly. The legal teams are ready, the client is waiting, and the bank deploys this state-of-the-art artificial intelligence to just, you know, do one final sweep of the compliance paperwork. Trevor Burrus, Jr.
SPEAKER_01Which sounds like a smart move.
SPEAKER_00Right. But instead of speeding things up, the AI confidently fabricates a critical regulatory violation, like one that doesn't actually exist at all.
SPEAKER_01Aaron Powell Oh, wow. A hallucination.
SPEAKER_00Yeah, complete hallucination. And it freezes the entire deal in its tracks, the client is panicking, and it forces like a dozen human analysts to spend their entire weekend untangling the mess.
SPEAKER_01Aaron Powell Yeah. And the stakes in that kind of environment are completely unforgiving. I mean, a generative AI hallucinating a historical fact in a uh a high school essay is just a quirky glitch. Right. But that same underlying tech misinterpreting a complex commercial credit risk profile that can trigger millions of dollars in fines. It can completely shatter institutional trust.
SPEAKER_00Aaron Powell And that harsh reality is really the core mission of today's deep dive. We are trying to cut through the endless hype surrounding AI to look at what is actually surviving in the trenches of these highly regulated, immensely complex industries.
SPEAKER_01Because there's a lot of noise out there right now.
SPEAKER_00So much noise. Right. We want to know are we anywhere close to a world where a bank can utilize AI to fundamentally anticipate your financial needs rather than just, you know, reacting to your requests?
SPEAKER_01And to answer that, we have a deeply revealing primary source today.
SPEAKER_00Yes, we do. It's an April 2026 interview from the IBSI FinTech Journal, and it features Sean Desmond, who is the CEO of Encino.
SPEAKER_01Yeah, and Desmond provides a really rare vantage point here. I mean, he took over the CEO role in February 2025, but his perspective is anchored in, well, over a decade serving as Encino's chief product and chief customer success officer.
SPEAKER_00He's been in the weeds.
SPEAKER_01Exactly. He is not a newly hired executive just dealing in abstract futuristic theories. He spent the last 10 years intimately dealing with the uh the mechanical nightmares of legacy banking technology. Trevor Burrus, Jr.
SPEAKER_00Trying to build actual pathways to modern digital architectures. Right. I do have to say, the publication paired his interview with some incredibly familiar visuals. You know the ones.
SPEAKER_01Oh, the classic fintech stock photos.
SPEAKER_00Yeah. A pair of disembodied hands holding a sleek tablet, glowing blue abstract code hovering in the air, and like a generic 3D pie chart labeled financial report in stark, bold letters.
SPEAKER_01Aaron Powell Right. It's the exact same visual vocabulary the industry has relied on for a decade.
SPEAKER_00It really is. The irony, though, is that while the stock imagery remains totally stagnant, the actual technology operating behind those glowing screens has undergone a violent fundamental paradigm shift.
SPEAKER_01Aaron Powell It really is a massive shift and it's architectural. I mean, the industry is rapidly abandoning the old paradigm where software was essentially just a digital filing cabinet.
SPEAKER_00Right, a passive system.
SPEAKER_01Exactly. A system that simply stored information until a human being arrived to retrieve it. We are moving towards systems where the software actively participates in the execution of the workflow.
SPEAKER_00Okay, let's unpack this. Because to understand where banking is going, Desmond insists we have to look at his core vision, which is a massive structural shift in who or what is actually doing the heavy lifting inside a financial institution. Right. He refers to this as the dual workforce.
SPEAKER_01Yeah, the theory driving Encino's 2026 strategy is to integrate what they call agentic AI into the hands of every single banker operating on their platform.
SPEAKER_00Agentic AI.
SPEAKER_01Right. They introduced these as digital partners late in 2025, and it represents a severe departure from the AI tools most consumers are currently familiar with. It absolutely is.
SPEAKER_00The best way to visualize the difference between traditional AI and agentic AI is to think about access and permissions. Traditional AI operates like a brilliant intern who happens to be locked inside a windowless room.
SPEAKER_01That's a great way to put it.
SPEAKER_00Right. You slide a prompt under the door, maybe ask for a client's debt-to-income ratio. The intern does the math flawlessly, writes it on piece of paper, and slides it back under the door.
SPEAKER_01Right. Highly intelligent but completely passive.
SPEAKER_00Exactly. It relies entirely on you to be the messenger and the initiator.
SPEAKER_01And an agentic AI fundamentally changes that dynamic by essentially holding the digital keys to the building.
SPEAKER_00Yes. The agentic AI has API permissions. It is operating like a highly proactive chief of staff. You don't have to slide a prompt under the door anymore. Trevor Burrus, Jr.
SPEAKER_01Right. Because within tightly defined pre-approved parameters, this AI can virtually walk down the hallway to the risk department servers.
SPEAKER_00Pull a specific file.
SPEAKER_01Cross-reference it with the compliance database. Notice that a secondary collateral document is missing and automatically draft an email to the client requipting that specific document.
SPEAKER_00And all of that happens before you've even poured your morning coffee. It operates autonomously to push the workflow forward.
SPEAKER_01What's fascinating here is how Desmond maps out the division of labor within this dual workforce concept. I mean, every single human banking role is assigned a corresponding digital counterpart, but they aren't duplicating effort.
SPEAKER_00Right. They're not doing the same job.
SPEAKER_01No, the design intentionally bifurcates the work based on cognitive strengths.
SPEAKER_00So the human banker and the digital partner sit side by side.
SPEAKER_01Exactly. And the digital partner absorbs the brutal high-volume analytical slog. We're talking about the massive data processing tasks, the cross-referencing of thousand-page documents, extracting unstructured data.
SPEAKER_00All the stuff that humans honestly hate doing anyway.
SPEAKER_01Right. And by entirely offloading that computational grind, the architecture frees up the human banker's cognitive bandwidth.
SPEAKER_00So they can't focus exclusively on high-level judgment.
SPEAKER_01Strategic decision making and navigating complex human relationships. The system does not replace human expertise, it acts as a force multiplier for it.
SPEAKER_00Here's where it gets really interesting. Because if this dual workforce model is so incredibly efficient, if it acts as the ultimate force multiplier, I'm honestly struggling to understand that MIT statistic we discussed at the top of the show.
SPEAKER_01The failure rate.
SPEAKER_00Yeah. If MIT researchers are proving that 95% of AI pilots in banking end in total failure, how does anyone actually code an AI to understand the archaic, convoluted mess of global banking?
SPEAKER_01Well, the failure rarely stems from a lack of computational intelligence. I mean, a generic large language model is trained on a massive corpus of data from the open internet.
SPEAKER_00So it knows what a loan is.
SPEAKER_01It fundamentally understands what a loan is conceptually. The breakdown happens entirely due to the environment.
SPEAKER_00Because the AI is basically stepping onto an alien planet.
SPEAKER_01That is the perfect way to look at it. Consider the digital reality of a traditional Western bank. You are often dealing with core infrastructure built on 40-year-old COBOL mainframes. Oh wow. And you have these intricately layered bespoke credit risk frameworks that vary wildly between jurisdictions. When a generic AI is dropped into that production environment, it encounters proprietary data structures it has never seen before.
SPEAKER_00It doesn't speak the local language.
SPEAKER_01Exactly. It tries to interpret a highly specific internal risk code, misaligns the context, and suddenly it's flagging a perfectly legal commercial transaction as a severe money laundering violation. Yikes. The pilot fails because the AI lacks native fluency in the bank's specific, highly regulated ecosystem.
SPEAKER_00I see the problem. So the sales team shows the bank executives this pristine, beautifully controlled demo where all the data is perfectly formatted and it looks like magic.
SPEAKER_01Yeah, a flawless tech demo.
SPEAKER_00But then they plug it into their actual 1980s mainframe. The AI panics at the messy data, hallucinates, and the whole thing catches fire.
SPEAKER_01That is exactly what happens.
SPEAKER_00So how does Encino claim to bypass this 95% failure rate? Like what is fundamentally different about their engineering?
SPEAKER_01The distinction really lies in their data architecture. Desmond is very clear that Encino did not approach AI as a flashy standalone application. Right. They didn't just purchase a generic LLM wrapper and bolt it onto the top of their existing software stack to satisfy impatient shareholders.
SPEAKER_00Aaron Powell So it's not just a chat widget hovering in the corner of the banker's screen asking, how can I help you today?
SPEAKER_01Far from it. The agentic AI is woven deeply into the structural fabric of the workflows that bankers already utilize every single day.
SPEAKER_00Aaron Ross Powell It's native to the system.
SPEAKER_01Exactly. And more critically, the AI's underlying models have been trained extensively on over a decade's worth of banking-specific financial services data.
SPEAKER_00Oh, I see.
SPEAKER_01It was architected from the very beginning to natively understand strict compliance mandates and complex audit requirements.
SPEAKER_00So the AI isn't an outside consultant trying to learn the bank's language. It was raised inside the bank.
SPEAKER_01That's it. When a relationship manager utilizes an NCO digital partner, they do not have to migrate to a new, unfamiliar platform.
SPEAKER_00Aaron Powell And just stay where they are.
SPEAKER_01Right. The AI inherently understands the specific compliance parameters of the loan there structure.