EP1000: Oracle sees the future…

This interview with IBSi FinTech Journal explores how artificial intelligence and cloud-native technologies are transforming the global banking industry. Oracle executive Venky Srinivasan explains that banks are evolving from traditional money custodians into data custodians, where embedded AI plays a major role in improving customer experiences, strengthening regulatory compliance, and detecting financial fraud more efficiently. The discussion highlights how many financial institutions are replacing large-scale transformation projects with progressive modernization, allowing faster and more flexible digital upgrades. Srinivasan also notes that both retail and corporate banking are prioritizing real-time and immersive banking experiences to meet growing customer expectations. Overall, the interview emphasizes that overcoming outdated legacy systems is critical for banks to remain competitive in the era of Open Banking, automation, and intelligent financial services.

Transcript

SPEAKER_00 0:00

Right now, uh the software holding roughly 10% of the entire global bank population's wealth, it's being quietly completely rewired by artificial intelligence. Aaron Powell Yeah.

SPEAKER_01 0:12

And we are talking about systems powering over 1,400 banks, I mean, across 160 plus countries. So it's massive.

SPEAKER_00 0:19

That is just staggering when you think about it.

SPEAKER_01 0:21

Aaron Powell Right. Because this isn't, you know, a conversation about some shiny new app update. It is literally a fundamental restructuring of the global financial nervous system. I mean, the decisions being made at this scale, they dictate how money moves around the whole planet.

SPEAKER_00 0:36

Aaron Powell, which is exactly our mission for this deep dive. Welcome everyone. Today we are exploring how modern banking is shifting from this uh static storage unit for your cash into a highly proactive data-driven engine, one that actually integrates directly into your lifestyle.

SPEAKER_01 0:52

Exactly.

SPEAKER_00 0:52

And our roadmap for this comes from a really fascinating June 2025 interview in the IBSI FinTech Journal. It's an interview with Venki Shannavasa.

SPEAKER_01 1:00

Yeah, he's the senior vice president and global head of sales for banking and insurance at Oracle Financial Services.

SPEAKER_00 1:05

Which, not coincidentally, happens to be the quiet giant behind that staggering 10% statistic we just mentioned.

SPEAKER_01 1:12

Right. They are the ones building this architecture.

SPEAKER_00 1:14

Okay, let's unpack this. Because to really see how this rewiring actually plays out, you know, we have to look at the raw data banks are sitting on and how they are changing their entire relationship to it. Aaron Powell Yeah.

SPEAKER_01 1:25

Well, historically, banks were kind of the ultimate passive containers, right? You deposit something valuable, they lock it in a ledger, and boom, the transaction is over.

SPEAKER_00 1:33

It just sits there.

SPEAKER_01 1:34

Aaron Powell Exactly. The data they held was purely historical, just a record of the past.

SPEAKER_00 1:38

Trevor Burrus, Jr.: And Srinavasan gives a really great hypothetical in the interview to kind of illustrate how this is changing. He uses a customer, uh, let's call him John Doe, who deposits a specific check on the 10th of every single month. Trevor Burrus, Jr.: Right.

SPEAKER_01 1:51

The monthly check example.

SPEAKER_00 1:53

Yeah. So for decades, that transaction was just a dead record in a database somewhere. But today, AI algorithms are ingesting like millions of these historical entries to establish dynamic behavioral baselines.

SPEAKER_01 2:06

Aaron Powell What's fascinating here is how the system maps John's behavioral velocity. It learns the pattern. Right. So if the tenth of the month passes and John just doesn't show up, the AI recognizes a disruption in that velocity. It flags the anomaly instantly. Wow. Yeah. And it actually prompts a human banker to pick up the phone, call John, and proactively check in.

SPEAKER_00 2:28

Which totally shifts the entire dynamic of the institution. I mean, we are so conditioned to banking being entirely reactive.

SPEAKER_01 2:36

Aaron Powell Oh, absolutely. You initiate a transfer, you apply for a loan.

SPEAKER_00 2:39

You complain about a random fee.

SPEAKER_01 2:41

Exactly. You have to make the first move.

SPEAKER_00 2:42

Aaron Powell Right. But here, the bank notices a broken pattern and reaches out to you. It feels, I don't know, less like a faceless corporation and much more like a hyper-observant hotel concierge.

SPEAKER_01 2:54

Aaron Powell That's a great way to put it.

SPEAKER_00 2:55

Aaron Powell You know, the kind who notices you didn't grab your usual morning coffee in the lobby, so they send someone up to your room to make sure you're all right.

SPEAKER_01 3:01

Aaron Powell Yeah, it's the inversion of causality. That's the key. The bank's data is no longer this passive storage thing. It's an active trigger for customer service. They're leveraging these incredibly granular data points to simulate a well, a highly personalized human level of care, but scaled across millions of accounts at the exact same time.

SPEAKER_00 3:23

But okay, there has to be a massive execution gap here, right?

SPEAKER_01 3:26

How so?

SPEAKER_00 3:27

Well, deploying AI sounds great in theory, but how are they actually building this into the system without it feeling like a cheap chat bot bolted onto the side of an old website? I mean, we've all interacted with those superficial AI overlays, and they're usually terrible.

SPEAKER_01 3:40

Well, they're the worst. But Oracle's entire philosophy, which they outlined pretty clearly in the interview, is designed to avoid that exact clunky add-on problem.

SPEAKER_00 3:49

So it's not just a plug-in.

SPEAKER_01 3:50

No, not at all. AI is not treated as a standalone product. You aren't just buying an AI module in a box and plugging it into your server room. The intelligence has to be fundamentally embedded directly into the applications across the full lifecycle of the banking products.

SPEAKER_00 4:04

Aaron Powell So it's the engine itself, not just a new coat of paint.

SPEAKER_01 4:07

Precisely.

SPEAKER_00 4:08

And the interview highlights a brilliant use case for this embedded AI in the world of compliance, actually.

SPEAKER_01 4:14

Right. The collection agents.

SPEAKER_00 4:16

Yes. Think about a bank's collection agents. Consumer protection regulators are incredibly strict about the exact phrasing those agents can use on a call with you.

SPEAKER_01 4:25

Yeah. One wrong word, and it's a massive fine.

SPEAKER_00 4:28

Exactly. And in the past, quality assurance just meant employing this army of humans to randomly sample recorded audio files days later, hoping to catch a mistake.

SPEAKER_01 4:38

Which is pretty much useless for actually preventing a compliance breach.

SPEAKER_00 4:42

Yeah.

SPEAKER_01 4:42

By the time the QA team hears the recording on a Tuesday, the damage was done on Thursday. Right. The regulatory fine is already triggered. But embedded AI changes that mechanism entirely. It uses natural language processing to analyze the live audio stream in real time.

SPEAKER_00 4:58

Wait, like as they are speaking?

SPEAKER_01 4:59

Yes. It converts the speech to text instantaneously, runs it against a dynamic database of all those regulatory frameworks, and flashes coaching prompts right on the agent's screen before they even finish their sentence.

SPEAKER_00 5:11

That is wild. It course corrects human behavior on the fly.

SPEAKER_01 5:15

It completely changes the margin of error for the bank. But you know, where the source really focuses this embedded AI power is on financial crime.

SPEAKER_00 5:25

Oh, right. The false positives. Because the traditional rules-based systems that banks have used for years to detect fraud, they generate a staggering 90% false positive rate.

SPEAKER_01 5:35

Aaron Powell 90%. It is an immense operational bottleneck. Because a traditional rigid system might have a hard-coded rule that just says, you know, flag any transaction over $10,000 originating from a foreign IP address.

SPEAKER_00 5:50

Okay. So a wealthy client buys a fancy watch while on vacation in Europe.

SPEAKER_01 5:54

And boom, the system instantly freezes the account.

SPEAKER_00 5:56

Right. And if 90% of your alerts are false, human investigators are essentially just doing data entry instead of actually investigating.

SPEAKER_01 6:04

Pretty much.

SPEAKER_00 6:04

It really reminds me of a building security guard whose monitors are hooked up to a motion sensor outside. But the sensor goes off every single time a stray loose blows across the perimeter.

SPEAKER_01 6:14

Right. So eventually the guard just starts hitting clear without even looking at the screen.

SPEAKER_00 6:18

Exactly. Which is exactly how actual threats slip through.

SPEAKER_01 6:22

If we connect this to the bigger picture, generative AI completely overhauls that entire workflow. The AI acts as the guard who actually checks the camera, recognizes the movement as just the wind, and suppresses the alarm automatically.

SPEAKER_00 6:35

So it filters the noise?

SPEAKER_01 6:36

Yes. The new models don't rely on simple binary rules anymore. They analyze deeper context to prioritize the alerts with the highest statistical probability of actual criminal activity.

SPEAKER_00 6:49

And it goes even further than that, actually, because when the AI does spot an actual threat, it doesn't just sound a smarter alarm, it builds an entire dossier. The source mentions the AI generates the alert narratives. How is it actually doing that synthesis?

SPEAKER_01 7:03

Well, it utilizes large language models that are trained specifically on financial topologies. So when a complex alert triggers, the model autonomously queries multiple internal APIs.

SPEAKER_00 7:14

Pulling from everywhere.

SPEAKER_01 7:15

Right. It retrieves contextual metadata, so transaction logs, email sentiment, swift messaging data and formats, all of that unstructured data into a comprehensive natural language report. Yeah. So the human investigator opens the file, and the first three hours of their work just gathering the context is already done.

SPEAKER_00 7:32

Okay, so the AI is embedded and it's doing this complex contextual synthesis. But here is where it gets really interesting for me because here's my issue.

SPEAKER_01 7:41

Okay, what is it?

SPEAKER_00 7:42

You can't just drop a state-of-the-art generative AI model onto a legacy banking mainframe from the 1980s. I mean, the processing power alone would melt the servers. How do these massive, highly regulated institutions actually upgrade the plumbing without taking the entire global financial system offline? Uptime is totally non-negotiable for them.

SPEAKER_01 8:03

You're absolutely right. They cannot pause global operations. Which is why Srinivasan points out that the ton line for these massive tech upgrades isn't actually shrinking.

SPEAKER_00 8:11

It's not.

SPEAKER_01 8:12

No. What's changing are the consumption patterns. He uses this great term, progressive modernization.

SPEAKER_00 8:17

Aaron Powell Progressive Modernization. Meaning they're finally abandoning those massive five-year rip and replace IT nightmares.

SPEAKER_01 8:25

Exactly. Because those monolithic projects always fail. By the time the bank finishes building the new core system, the underlying technology is already obsolete anyway. That makes sense. So progressive modernization means delivering iterative, modular pieces of business value.

SPEAKER_00 8:41

It's kind of like replacing the engines on a commercial airliner while it's cruising at 30,000 feet. Yes. You can't land the plane, but how do you actually swap out a core banking function without crashing the old legacy database that it relies on?

SPEAKER_01 8:56

Aaron Powell By fundamentally changing the software architecture to cloud native composable services. Think of it less like a tangled web of code and much more like a global shipping port.

SPEAKER_00 9:07

Aaron Powell Okay, let's follow that. How does the shipping port work for banking data?

SPEAKER_01 9:10

Aaron Powell Well, in a modern port, everything moves in standardized shipping containers, right? Sure. You can swap out a container full of electronics for a container full of grain. The massive crane moving them doesn't care what is inside. It just knows how to interface with the standardized corners of the box to move it from the ship to the truck. Trevor Burrus, Jr.

SPEAKER_00 9:27

Right. The corners are universal.

SPEAKER_01 9:28

Aaron Ross Powell Exactly. So in banking, the software is broken down into independent microservices. Those are your containers. And the APIs, the application programming interfaces, they act as the cranes, routing the data between them.

SPEAKER_00 9:40

Aaron Powell Oh, I see. So a bank isolates its, say, deposit processing microservice, they can build a brand new, highly efficient cloud version of that service while the old one is still up and running.

SPEAKER_01 9:52

Yep. And the APIs simply start routing the data to the new container. Once they verify it works seamlessly, they just decommission the old one. Zero downtime.

SPEAKER_00 10:00

That is brilliant.

SPEAKER_01 10:01

It really is. And this composable architecture is what enables what the industry calls straight through processing.

SPEAKER_00 10:07

Aaron Powell Explain the mechanism behind that. What exactly is flowing straight through?

SPEAKER_01 10:11

Aaron Powell The data. The transaction flows from initiation. Say you make a complex cash management request on your smartphone through the security checks into the ledger and achieves final settlement completely automatically.

SPEAKER_00 10:23

So no human bottlenecks.

SPEAKER_01 10:25

There are no manual data entry points, no human approvals waiting in the queue, and no batch processing delays at the end of the day.

SPEAKER_00 10:32

Okay. So we have progressive modernization pushing banking toward these hyper-efficient, composable cloud models. But the catalyst forcing this evolution isn't just, you know, a desire for better back-end tech. It's raw, unyielding customer expectation. Oh, absolutely. The source describes a human-centric view that highlights a pretty severe duality in the banking industry right now.

SPEAKER_01 10:55

Yeah, the immense friction between the consumer experience and the enterprise experience.

SPEAKER_00 11:00

Right. Because the executive who manages a complex multinational supply chain on Monday morning, that is the exact same human being who spent Sunday afternoon booking flights and ordering groceries on flawlessly intuitive apps.

SPEAKER_01 11:12

Exactly. Same brain, same expectations.

SPEAKER_00 11:14

And they bring those frictionless expectations right into the boardroom. Their tolerance for clunky, archaic corporate software is officially zero.

SPEAKER_01 11:23

And the stakes in corporate banking are astronomically high. You are dealing with global cash management, sweeping liquidity across dozens of currencies in real time, navigating intricate trade finance protocols. It's huge. Because of that immense complexity, corporate banking technology historically lagged years behind retail apps.

SPEAKER_00 11:44

But that lag is over. Sharina Vasan drops a statistic in the interview that I think perfectly encapsulates this convergence.

SPEAKER_01 11:51

The hundred days stat.

SPEAKER_00 11:52

Yes. He notes that executing corporate credit historically took up to a hundred days.

SPEAKER_01 11:57

Wow.

SPEAKER_00 11:57

A hundred days of manual underwriting, physical document verification, and endless approval cues.

SPEAKER_01 12:04

Over three months just to secure business funding? That's a lifetime in business.

SPEAKER_00 12:08

And today, using these composable AI architectures we've been talking about, that exact same corporate credit execution is instant.

SPEAKER_01 12:16

Instant. Because straight-through processing eliminates the cues entirely. The AI models instantly pull the corporation's historical cash flow data via those APIs, cross-reference it against real-time market risk parameters, and execute the credit decision autonomously.

SPEAKER_00 12:38

So the model is doing the heavy lifting.

SPEAKER_01 12:41

Exactly. The highly complex risk assessment that used to take a whole team of analysts three weeks to compile, it is performed by the model in the time it takes the screen to refresh.

SPEAKER_00 12:51

So what does this all mean? It means the enterprise tech world has nowhere left to hide.

SPEAKER_01 12:56

No, they don't.

SPEAKER_00 12:57

If a bank can underwrite a massive corporate loan instantly, the expectation for literally every other financial service becomes immediate execution. So if 100 days drops to zero, what happens to the entire financial ecosystem over the next few years?

SPEAKER_01 13:13

Well, we are looking at a landscape dominated by hyperscale technology platforms. Platforms designed to support true 2047 continuous settlement. The industry is rapidly moving away from that old batch processing at the end of the business day.

SPEAKER_00 13:27

Right, waiting for the bank to close.

SPEAKER_01 13:29

Exactly. It's moving to a reality where money and data move instantly, all the time. Over the weekend, holidays, doesn't matter. Trevor Burrus, Jr.

SPEAKER_00 13:36

And that fundamentally relies on open banking, right? And that API interoperability we were talking about earlier with the cranes and the shipping containers.

SPEAKER_01 13:42

Aaron Powell Yes. Those APIs become the digital connective tissue for everything. For corporate clients, this accelerates the rise of banking as a service or buys.

SPEAKER_00 13:52

Moss, right?

SPEAKER_01 13:53

Yeah. The bank is no longer just a destination portal that a CFO logs into. Instead, the bank's entire suite of services, cash management, real-time payments, tailored financing, it's embedded directly into the corporation's own enterprise resource planning software.

SPEAKER_00 14:09

So the banking infrastructure becomes completely invisible.

SPEAKER_01 14:13

Almost entirely. A massive retailer simply manages its money directly inside the same software it uses to run its daily supply chain without ever opening a separate banking application.

SPEAKER_00 14:23

That is incredible. And for the retail consumer, it becomes even more integrated into their lifestyle, right?

SPEAKER_01 14:28

Oh, definitely. Utilizing open banking, a primary financial institution, can see your holistic financial footprint across multiple accounts. The bank leverages generative AI to actually map your financial trajectory.

SPEAKER_00 14:40

Okay, give me an example of what that looks like. Sure.

SPEAKER_01 14:42

So it might notice you have a growing high yield savings account and also a recurring pattern of browsing real estate sites.

SPEAKER_00 14:48

Oh wow. So it proactively offers a customized, pre-approved mortgage structure before you even fill out an application.

SPEAKER_01 14:56

Exactly. It anticipates the life event. But there is a massive tension looming over the industry that Srinavasan touches on at the end of the interview.

SPEAKER_00 15:05

The democratization of finance.

SPEAKER_01 15:06

Yes. The giant traditional banks are facing an existential threat from that. Nimble, highly focused fintech startups are innovating at breakneck speeds.

SPEAKER_00 15:17

And those fintechs don't have to progressively modernize, do they? Because they were born in the cloud.

SPEAKER_01 15:21

That's the kicker. They aren't burdened by 40-year-old mainframes. They can hyperfocus on one specific, niche-like international remittances or peer-to-peer lending and just engineer a flawless, frictionless user experience.

SPEAKER_00 15:34

Aaron Powell, which is forcing the massive traditional banking giants into a corner. I mean, they have no choice but to rethink their operating models from the ground up and shift their sprawling infrastructures entirely to the cloud.

SPEAKER_01 15:45

Aaron Powell They have to. And Sri Dev Awesome warns that the ultimate battleground here isn't who has the slickest app interface.

SPEAKER_00 15:52

Right. The battleground is the data. Trevor Burrus, Jr.

SPEAKER_01 15:54

Always the data.

SPEAKER_00 15:54

Because if your generative AI is hallucinating or making bad predictions based on fragmented siloed data from a 1990s database, you are going to lose the customer instantly. Clean, highly granular, meticulously organized data is really the only way traditional institutions can survive this fintech invasion.

SPEAKER_01 16:15

Absolutely. A model is only as intelligent as the data it consumes. If the input is garbage, the AI is garbage.

SPEAKER_00 16:21

Which really brings us full circle. As we've seen today, our mental model of a bank just has to change. They are no longer the static vaults storing our physical currency.

SPEAKER_01 16:30

Right. Through embedded AI systems that actively monitor compliance and autonomously gather context for fraud investigations, they are fundamentally altering their core operations.

SPEAKER_00 16:39

And by utilizing that progressive modernization and composable cloud native architectures.

SPEAKER_01 16:44

Basically rebuilding the airplane while flying it.

SPEAKER_00 16:46

Exactly. They are blurring the lines between consumer ease and corporate complexity, turning hundred-day underrating processes into instant executions.

SPEAKER_01 16:55

It really is an invisible revolution. Every time you skip a visit to a physical branch or a payment clears instantaneously over a long weekend, you are riding on top of a fiercely competitive, globally connected AI infrastructure, one that is working tirelessly to analyze your behavior and anticipate your next move.

SPEAKER_00 17:16

Which leaves you with one final thought to mull over as we wrap up today. We started by talking about how banks have shifted from custodians of money to custodians of data. But if open banking and generative AI allow your financial institution to successfully anticipate your life goals, your career changes, and your spending habits before you even fully articulate them to yourself.

SPEAKER_01 17:36

That's a wild thought.

SPEAKER_00 17:37

At what point does your bank stop being a financial service provider and start becoming an active director of your life choices?

SPEAKER_01 17:44

Yeah. It is a horizon we are approaching much faster than people realize.

SPEAKER_00 17:48

It really is. Well, thank you so much for joining us on this deep dive. Keep asking questions, keep looking behind the curtain, and we'll catch you next time.