EP961: Rethinking Banking: The Future of AI in Financial Services

As artificial intelligence continues to make waves in the banking sector, this episode dives into how it’s reshaping working capital management. We discuss how AI, APIs, and automation are unlocking new efficiencies in liquidity, supply chain resilience, and sustainable finance. The conversation covers AI’s transformative role in improving credit scoring, fraud detection, and operational agility. With insights on AI-driven decision-making models, the episode also highlights the balance between speed and responsible risk management. As financial institutions scale globally, the need for innovation and regulatory adaptability becomes even more critical.
 
Discover how AI is not just enhancing but revolutionizing banking and financial services for the future.

Transcript

SPEAKER_01 0:00

Think about the last time you um interacted with your bank. Was it just this clunky, frustrating experience? Or did it actually feel, you know, strangely intuitive?

SPEAKER_02 0:10

Aaron Ross Powell Right, because usually we hold on to this image of banking as this rigid, slow-moving monolith just bound by red tape.

SPEAKER_01 0:17

Exactly. But today we are bringing you a deep dive into how banking is quietly undergoing this massive, AI-driven revolution right under our noses. I'm your host, by the way.

SPEAKER_02 0:28

Aaron Ross Powell And I'm here to help guide you through the technical side of this shift because the reality of what's happening inside those server rooms, it's radically different than what most people think.

SPEAKER_01 0:36

Aaron Powell It really is. So our source material for this deep dive is a really insightful October 2025 interview from the IBSI FinTech Journal. It features Ramesh Malia, who is the Chief Technology Officer of DBS Bank India. And the piece was written by Puja Sharma.

SPEAKER_02 0:52

Yeah, and it's such a revealing interview, mostly because it um it strips away that glossy corporate veneer and gets right into the actual mechanics of digital transformation.

SPEAKER_01 1:01

Right. We aren't just talking about shiny new app features here.

SPEAKER_02 1:03

We aren't just talking about shiny new app features here. No, not at all. We are looking at the foundational rewiring of global financial plumbing.

SPEAKER_01 1:12

Oh, and to set the mood for where this tech is heading, the journal actually includes this incredible visual context. Picture a professional interacting with a futuristic, glowing blue, holographic dashboard.

SPEAKER_02 1:24

Oh, yeah, that image is striking.

SPEAKER_01 1:26

Right. Right in the center, there is an icon of a human head with a gear turning inside, and it's surrounded by these complex data streams and the words natural language processing.

SPEAKER_02 1:36

It looks like a scene pulled straight from a sci-fi film.

SPEAKER_01 1:39

It really does. But it represents the very real, highly sophisticated neural networks being deployed right now to handle your money. So our mission today is to unpack exactly how these massive financial institutions are rewiring their ancient digital plumbing with cutting-edge AI, open banking, and cloud tech.

SPEAKER_02 1:58

And doing all of that without crashing the systems that hold our money, which is the really tricky part.

SPEAKER_01 2:03

Seriously. Okay, let's unpack this. I want to start with the first major theme of the interview, which is AI on the front lines, but also the heavier lifting happening behind the scenes.

SPEAKER_02 2:13

Yeah, the core idea the CTO drives home is that AI isn't just some buzzword for them. It's an enabler of digital transformation. And it's measured strictly by customer impact and business value.

SPEAKER_01 2:24

Aaron Powell So they are actively moving past that whole hype cycle.

SPEAKER_02 2:28

Aaron Powell Exactly. Like, for instance, the deployment of AI-powered virtual assistants. It isn't just about offering a neat chat interface, it's about fundamentally driving up customer self-service rates.

SPEAKER_00 2:39

Right.

SPEAKER_02 2:39

And when you combine that with predictive analytics, you shift from reactive banking to proactive banking.

SPEAKER_01 2:44

Aaron Powell Wait, what does that actually look like for the consumer?

SPEAKER_02 2:47

Aaron Powell Well, the system analyzes your behavioral patterns to deliver tailored financial insights directly to you, the customer, often before you even realize you need them.

SPEAKER_01 2:55

Oh, wow. But the truly heavy lifting is happening in the credit departments, right? The interview notes they are using AI-driven models to process credit applications. Aaron Powell Yeah.

SPEAKER_02 3:04

And that actually significantly reduces non-performing assets. Because legacy underwriting models rely heavily on static historical data. Things like bureau scores or past tax returns. Trevor Burrus, Jr.

SPEAKER_01 3:17

List stuff that takes forever to gather.

SPEAKER_02 3:18

Trevor Burrus Exactly. Trevor Burrus But an AI-driven model ingests massive real-time streams of alternative data. Trevor Burrus Yeah. You can look at a small business's daily supply chain health or their digital payment flows.

SPEAKER_01 3:30

Aaron Ross Powell So it's much faster and way more accurate. But um I have to play the skeptic for a second.

SPEAKER_02 3:35

Aaron Powell Go for it.

SPEAKER_01 3:36

Trevor Burrus They mentioned this internal initiative that caught my attention. They launched something called DBS GPT, which is a generative AI tool for handling documents and queries.

SPEAKER_02 3:45

Aaron Powell Yes, a very powerful internal tool.

SPEAKER_01 3:47

Aaron Powell Right. But when I hear a bank is launching its own DBS GPT to handle tasks, my immediate thought is, you know, they are just trying to replace human jobs with robots to save a buck.

SPEAKER_02 3:59

Aaron Powell I mean that is the natural assumption. But the actual goal here isn't replacement, it's reallocation.

SPEAKER_01 4:04

Aaron Ross Powell Okay, how so?

SPEAKER_02 4:06

Aaron Ross Powell Think about the sheer volume of unstructured data in a bank. Right. Complex commercial loan originations, multi-layered KYC documents. Historically, extracting data from those has required armies of human analysts just doing tedious manual verification.

SPEAKER_01 4:20

Aaron Powell Oh, I can only imagine just staring at forms all day.

SPEAKER_02 4:23

Aaron Powell Right. It's exhausting. So DBSGQT handles that tedious document work. It offloads the rote mechanics to the machine.

SPEAKER_01 4:32

Aaron Ross Powell So that the human employees are freed up.

SPEAKER_02 4:34

Exactly. They are freed up to focus on higher value, personalized advisory, and strategic planning. They review the AI's output in minutes and then spend their time actually sitting down with a client.

SPEAKER_01 4:45

Oh, I see. Which actually ties perfectly into the subtitle of the article itself. It was rethinking banking when AI learns to care.

SPEAKER_02 4:54

Yes. The AI handles the logic parsing so the humans actually have the bandwidth to care about the customer.

SPEAKER_01 4:59

Aaron Powell Okay, that makes sense. But here is the massive technical contradiction that I'm stuck on. How do you plug lightning fast, futuristic AI into the ancient monolithic legacy software that most banks are built on?

SPEAKER_02 5:14

Ah, yeah. That transition into legacy modernization is probably the most terrifying engineering challenge any massive institution faces today.

SPEAKER_01 5:21

Aaron Powell Right. Because people often use that Jenga Tower metaphor for legacy code. Like you can't just kick the table over and start fresh when people's life savings are on the line.

SPEAKER_02 5:29

Exactly. It'd crash the whole system.

SPEAKER_01 5:30

You have to carefully swap out one wooden block at a time while the tower is still standing. Or honestly, it's more like trying to change a tire on a moving bullet train.

SPEAKER_02 5:40

That bullet train analogy captures the operational terror perfectly because you have this colossal architecture processing millions of transactions.

SPEAKER_01 5:49

So how do they actually do it without derailing everything?

SPEAKER_02 5:52

According to the CTO, DBS takes a quote, prudent, multi-pronged approach. They have to balance innovation with absolute continuity.

SPEAKER_01 6:02

Okay, what does that look like practically?

SPEAKER_02 6:04

Well, they are taking these massive monolithic systems and breaking them down into what are called microservices.

SPEAKER_01 6:10

Microservices. So instead of one giant system where updating the mortgage software accidentally crashes the ATM network.

SPEAKER_02 6:17

Yes, exactly. Microservices isolate those functions. So account balances, payment gateways, and loan processing become independent decoupled components.

SPEAKER_01 6:26

Wow, okay. So they can replace or replatform individual parts while keeping the core operations totally stable.

SPEAKER_02 6:33

Precisely. And to support that, they rely heavily on hybrid cloud environments, gives them massive scalability and much shorter deployment cycles.

SPEAKER_01 6:41

But doing that on the fly must require incredible discipline.

SPEAKER_02 6:44

Oh, absolutely. They rely on strong DevOps, automated testing, and strict rollback strategies. So every change is made with confidence.

SPEAKER_01 6:52

Wait, how does a rollback strategy work if something goes wrong?

SPEAKER_02 6:55

Let's say they push an update to a microservice, but they only route 1% of live traffic to it. If the testing framework detects any microscopic bug or performance drop, it instantly triggers a rollback.

SPEAKER_00 7:06

Instantly.

SPEAKER_02 7:06

In milliseconds. It routes the traffic back to the older, stable version. The customer never even sees an error screen.

SPEAKER_01 7:13

That is wild. Okay, so once you break the system into smaller blocks, how do the old untouchable core systems communicate with the shiny new cloud microservices?

SPEAKER_02 7:22

Aaron Powell That is where the API first model comes in, which was highlighted in the text. They build APIs to act as modern interfaces.

SPEAKER_01 7:28

Like adapters.

SPEAKER_02 7:30

Exactly like adapters. They wrap around the old legacy functions. It protects the critical older systems from unnecessary changes, but lets them talk seamlessly to the newer cloud services.

SPEAKER_01 7:40

Aaron Powell Okay, so that leads us right into the next section. Because once you've built these secure API adapters to make your internal systems talk to each other, you suddenly realize you can use those exact same adapters to let outside innovators plug into your bank.

SPEAKER_02 7:53

Yes. And that represents a massive shift. Open banking and API ecosystems are opening huge new avenues for innovation and revenue.

SPEAKER_01 8:02

The text mentions that these open APIs allow the bank to embed financial services directly into a customer's journey on digital platforms.

SPEAKER_02 8:10

Right, like embedding a loan approval right into a third-party checkout screen. By partnering with Agile FinTechs, the bank can diversify its products and reach new segments much faster.

SPEAKER_01 8:20

Okay, but wait, I have to challenge this open door policy? Aren't fintechs the scrappy startups trying to literally steal the big bank's lunch?

SPEAKER_02 8:27

That's how it used to be viewed, sure.

SPEAKER_01 8:29

So why would DBS Bank give their competitors an open API to plug right into their own ecosystem?

SPEAKER_02 8:35

Well, what's fascinating here is the symbiotic relationship outlined in the text. It's not a zero-sum competition anymore.

SPEAKER_01 8:42

Really? How so?

SPEAKER_02 8:43

Because the fintechs have incredible agility and speed. They could build phenomenal user interfaces. But what they lack is what the traditional bank has massive scale, deep trust, and a century of regulatory expertise.

SPEAKER_01 8:57

Oh, I see. So it's a co-creation strategy.

SPEAKER_02 8:59

Exactly. The fintech handles the slick interface, and the bank handles the heavy regulatory lifting and capital backend. It benefits you, the listener, by bringing the best of both worlds to your banking app.

SPEAKER_01 9:11

Okay, that makes total sense. However, if you are breaking your system into microservices, operating on a hybrid cloud, and opening your doors to third-party fintech startups, haven't you just exponentially increased your vulnerability to hackers?

SPEAKER_02 9:24

You absolutely have. You've multiplied the attack surface by an order of magnitude.

SPEAKER_01 9:28

Which sounds terrifying.

SPEAKER_02 9:29

It does. And that is exactly why operational resilience in Fort Knox 2.0 level cybersecurity become non-negotiable.

SPEAKER_01 9:37

Right. The CTA actually states they adopt a security by design philosophy. So security isn't just a patch applied at the end, it's baked into the initial blueprints from the start of any build.

SPEAKER_02 9:48

Precisely. They are using advanced AI-driven monitoring for real-time threat detection. An AI model analyzes the telemetry of every single API call, looking for microscopic anomalies.

SPEAKER_01 10:00

Aaron Powell And they pair that with regular penetration testing, right? Constantly attacking their own systems to find vulnerabilities.

SPEAKER_02 10:07

Yes. But resilience is also about keeping the lights on. It means high availability architecture, rigorous disaster recovery planning, and massive redundancy.

SPEAKER_01 10:16

Aaron Powell So that the services stay up even under challenging conditions?

SPEAKER_02 10:20

Exactly. And all of this rapid tech adoption is strictly governed by an ethical structure. They call it the PRE framework.

SPEAKER_01 10:26

Oh yeah, the PRV framework. This stood for purposeful, unsurprising, respectful, and explainable.

SPEAKER_02 10:32

That's the one.

SPEAKER_01 10:32

Now here's where it gets really interesting to me. Purposeful, respectful, explainable, those make sense, but unsurprising.

SPEAKER_02 10:40

It stands out, doesn't it?

SPEAKER_01 10:41

Yeah. Because in the tech world, every company wants to like surprise and delight you. Why is a major bank explicitly making it their ethical framework to be unsurprising?

SPEAKER_02 10:51

Aaron Powell Well, let's contextualize this for you, the listener. When it comes to how a machine learning algorithm handles your private financial data, surprise is the absolute last thing you want.

SPEAKER_00 11:02

Right. A surprise in my bank account usually means panic.

SPEAKER_02 11:04

Exactly. You want transparency and predictability. When dealing with stochastic models, they are designed to be creative. But you don't want a creative algorithm assessing your mortgage application.

SPEAKER_00 11:16

Definitely not.

SPEAKER_02 11:17

So this ethical framework ensures their tech remains grounded in customer trust. The AI must be logical, consistent, and unsurprising.

SPEAKER_01 11:25

Okay, so with this hyper-secure, highly resilient, and ethically governed foundation finally in place, what does the CTO say is coming next for your banking experience?

SPEAKER_02 11:35

Looking ahead, the focus shifts heavily to hyperpersonalization, zero latency, and data governance. There are three massive tech leaps driving this. Let's hear them. Hyper automation, generative AI, and edge computing.

SPEAKER_01 11:47

Okay, let's break those down. The text talks about hyperautomation combining AI and machine learning with robotic process automation or RPA.

SPEAKER_02 11:56

Right. Essentially deploys software bots to eliminate manual intervention across complex workflows.

SPEAKER_01 12:02

And then generative AI moves past those annoying traditional chat bots, right? Sure. Into highly personalized experiences that actually anticipate your needs in real time.

SPEAKER_02 12:12

Aaron Powell Exactly, which ties right back to that futuristic natural language processing dashboard image we talked about.

SPEAKER_01 12:17

Aaron Powell Right, the glowing blue hologram. But to achieve that real-time anticipation, the text brings up edge computing, which means processing data closer to the source.

SPEAKER_02 12:27

Yes, as connected devices proliferate.

SPEAKER_01 12:29

But I need clarification here. Why does processing data closer to the source actually matter for a consumer using a banking app on their phone?

SPEAKER_02 12:37

Aaron Powell Well, if we connect this to the bigger picture, it changes how your data moves. Traditionally, your raw data travels from your phone to a centralized server hundreds of miles away, gets processed, and comes back.

SPEAKER_01 12:48

Which creates a delay, right?

SPEAKER_02 12:49

Aaron Powell Yes, network latency. But processing at the edge, like right on the neural chip of your device, enables ultra low latency services, meaning instant reactions.

SPEAKER_01 13:00

Ah, so the app is lightning fast.

SPEAKER_02 13:02

Yes. But crucially, it strengthens data privacy. Because your raw data doesn't have to constantly travel back and forth over vulnerable networks. It stays on your device. Everything circles back to data governance.

SPEAKER_01 13:14

Wow. Okay, so what does this all mean for you, the listener? We've gone from internal DBS GPTs handling tedious documents to the delicate Jenga game of migrating to microservices without crashing the system.

SPEAKER_02 13:27

Yep, and out into the open API ecosystem where banks and fintechs co-create.

SPEAKER_01 13:31

Right. And finally into this hyper-automated edge computed future. The overriding theme really is best summarized by Ramesh Malia's own quote from the piece. Oh, I love this quote. Yeah. He says by embedding AI and emerging technologies responsibly, we're not just automating processes, we're rethinking how customers and employees experience banking every day.

SPEAKER_02 13:50

There's a profound shift.

SPEAKER_01 13:52

It really is. It means your banking experience is about to become invisible, instantaneous, and highly personalized, all without you having to sacrifice your privacy or security to get it.

SPEAKER_02 14:02

Which is incredible. Yep. But you know, it does leave us with one highly provocative thought to ponder.

SPEAKER_00 14:08

Oh, laid on me.

SPEAKER_02 14:10

If hyperautomation and generative AI reach a point where they can flawlessly anticipate our financial needs and proactively manage our wealth in real time, completely in the background, what happens to our own financial literacy?

SPEAKER_00 14:23

Oh, wow.

SPEAKER_02 14:24

Right. In a future where the AI learns to, quote, care for our money entirely behind the scenes, do we risk forgetting how to care for it ourselves? Something to think about the next time your financial app magically executes exactly what you needed before you even knew you needed it.