Cameron Smonk (00:18)
Welcome to Banking Blueprints. I’m Cameron Smonk and today I’m joined with Lucy Li, Financial Services Industry Advisor at Microsoft, and Mun Tham, Industry Advisor at Zafin. To kick us off, I want to ask you about your experiences in finance and technology—maybe a surprising moment within your journeys.
Mun Tham (00:35)
Sure, I can get it started. So I was actually an engineer by training, but I immigrated to Canada back in 2007, right before the subprime crisis. So I came here as an engineer, lost my job right after, but that whole experience got me really interested in banking—financial services—who are the central bankers.
So I went back to school, got my master’s, and started working in banks for 15 years thereafter. So from that experience, I learned a few things. Banks are great at making money, but they’re not so good at innovating, which is why ultimately I’m now with Zafin—where, as a fintech, we can really drive the change, implement all these new strategies and technology in the banking space, and help banks innovate. What about you, Lucy? How was your path?
Lucy Li (01:05)
I’ve always been in banking and consulting—I think for 17 years before joining Microsoft. And I think just before joining Microsoft, I was already picking up in commercial corporate banking where I was in, there wasn’t a ton of product innovation you could do on the core banking products, especially on the corporate side. And really a lot of the change for innovation is going to come from the digital experience—kind of how the bank delivers these services to their clients.
So I was very interested in a digital banking path and I wanted to explore more on the frontier technology in banking. So I was able to get this very interesting financial advisory role in Microsoft, and I think what the unexpected turn in this was more because ChatGPT was launched very shortly after I joined Microsoft. I think within maybe 12 months, Microsoft also launched our own enterprise M365 Copilot tool, and we as advisors are expected to know how our clients—the banks—are using these tools. So I got to experiment with these GenAI tools very quickly and do a very deep dive into generative AI. So this was just perfect timing, but very unexpected.
Cameron Smonk (02:49)
Well, that’s actually a great segue because in this AI-driven new horizon, things are changing pretty fast. So I wanted to ask: how are digitization, AI, evolving client expectations, and regulatory pressures reshaping corporate banking today?
Mun Tham (03:06)
Okay, I can get us started. Let’s talk about client experience first, right? When you think about the corporate clients—the CFOs, the corporate treasurers—personally, on the personal front, they’re getting all this innovative experience, you know, from the app, from the phone—everything’s at their fingertips. Naturally, the expectation when it comes to corporate banks is, you know, information is right there at the right time, fully transparent. And they almost expect everything to be curated—to have context that’s relevant to them.
So that’s one: client experience has definitely changed. Another would be regulatory pressure that’s driving a deep dive on product economics. So we are hearing from banks that the bank boards are really questioning the profitability of each client—are they making money after all the discounts and all the bespoke deals have been done? So we are seeing the pressure to really drive relationship economics—that granularity.
The third one will be technology shift. We see so many cloud data programs, AI pilots, and real-time projects. So corporate bankers are always trying to cope with all this landscape kind of change. But at the same time, while the banks want to innovate, they are trapped because at the very core level, product pricing and billing information are still built on core systems that were launched in the 70s or the 80s. Until they digitize that, centralize it, it’s very hard for them to innovate. They will always come back to this bottleneck—and that’s where Zafin comes in.
So as a company, we are core agnostic. We are basically the platform to centralize product pricing and billing information for banks. We sit on top of the core. We are the brain that controls the economics of banking activities.
So that’s how I see the trends happening and what Zafin can do in this space. What about you, Lucy—what is Microsoft doing?
Lucy Li (05:12)
Well, I think building on what you are seeing, we also see the same trends impacting and triggering the reactions of the corporate banks that we interact with. So I get a lot more questions—especially in the last 12 months—from our clients saying, “This must be something AI can solve for. We must be able to automate this process with AI.” They’re questioning: why are we doing things manually?
I think it’s going to change expectations on the banker side. They don’t want to rely on their data analytics team to deliver client insights. They want all the client data and internal data unified on the same platform. They want to be able to query on their own and create something customized—what they need to know about the client before they interact with the client. And so it’s very much like an on-demand, more liberated access for bankers to understand what clients may have with the bank, what products or solutions the banks want to promote, and how they can provide more personalized advice.
I’ll also add: on the corporate treasurer side, there is changing demand because treasurers and finance users in F&L are being exposed to generative AI tools too. So they’re saying, “Well, if I’m able to interact through conversational UI with my data, my banks—if they’re going to embed into my environment—should be able to do that as well. Give me a different experience.”
So I do think corporate treasurers are expecting a very different interaction with banks. If you look to maybe five, ten years down, embedded finance is going to be completely different. It’s no longer just initiating payments in the corporate environment. It’s more about how banks would communicate with corporate treasurers through some sort of agentic AI interaction—and the treasurer is able to pull data and complete actions using an AI agent. And then the interaction will be less about what the bank’s online transaction banking platform can do versus what the treasurer is expected to do in their own environment.
Mun Tham (07:57)
Yeah—like embedded finance and agentic flow, that’s what’s going to happen.
Cameron Smonk (08:02)
Yeah—I’d love to ask a question of you both. What are the similarities and differences in how Microsoft and Zafin approach co-innovation? Mun, do you want to start?
Mun Tham (08:12)
Actually, Lucy, do you want to go first? I would like to hear what Microsoft is doing first.
Lucy Li (08:18)
So we do a lot of innovation. But I would say among the many flavors of innovation that we do with clients—especially corporate clients—there are two main categories. One is when we tackle existing patterns, but apply them to new use cases in banking. An example of this is: we know agentic AI has the ability to automate the workflow of ingesting information, creating summaries, and providing analysis. And you can take that pattern and put it in the credit underwriting process—so existing pattern, perhaps new use cases.
The second type of innovation is truly co-innovating with clients. This means we get into new patterns and new capabilities, and that’s when our product group gets involved. One example of this is agentic commerce. We’re actively working with payment processing companies, very large corporates, and banks looking at how payments would happen in conversational AI chatbots—including ChatGPT and Microsoft’s consumer Copilot. These inform our product roadmap.
I’ll give an example of an innovation we completed recently: a trade finance proof of concept. We announced this at Sibos where we worked with the International Chamber of Commerce, ANZ, Lloyds Bank, and HSBC Bank to digitize documentary trade end-to-end—from the source system (ERP) all the way to banks ingesting structured datasets. We consider this an existing capability applied to a new use case.
So what we did: we used Dynamics 365 ERP. We looked at how an exporter could take a letter of credit transaction and all the required documents and data elements and, instead of sending paper documents, send structured data from their ERP packaged into structured data files. This is done with an AI agent. The AI agent gets instruction from the bank about required documents, triggers an alert to the exporter, and in the exporter’s environment the treasurer instructs the AI agent to start pulling the data. The AI agent finds the data elements in the ERP, creates the data files, and validates discrepancies across documents—like buyer name or buyer address appearing differently.
Then the package is ready to send to the bank. In this case, the bank can ingest structured data directly—no OCR, no manual entry. This shows how we bring the latest GenAI technology to ANZ, HSBC, and Lloyds and test whether we can eliminate paper documents while improving time savings and client experience. And what’s unique about Microsoft in co-innovation is that we stretch the capability of the design—we believe it’s not just about connectivity, but embedding agentic workflow into connectivity. That’s a good example of how we co-innovate.
Mun Tham (13:04)
The way you talk about it is from the corporate client perspective, so I would follow on to that from the bank’s perspective. If the client is using agent AI flows, the bank will need to make sure product pricing and billing information is agent-ready.
So back to our main value prop at Zafin: we become the economics layer of the bank. This layer keeps the product information, the pricing information, and the billing information—so we centralize it, digitize it, and make it agent-ready.
What it means for us: when we work with clients, we need to understand how far along they are and how far they want to innovate. First, have they centralized economics information? If not, that’s what we do with them. After that, if they have enough product information in the Zafin layer, we can deploy AI to drive pricing optimization—understanding client value, taking the entire relationship into consideration when pricing a deal to support the bank’s sales team.
And finally, if the end client has agentic workflow, we make sure the information is machine-readable so the bank can embed services into the client’s ERP system. That’s how we think about innovating with our clients. We do this through our client advisory board, banking leadership summit, and ongoing client conversations. Unlike big tech like Microsoft, we take a more personal approach.
Cameron Smonk (14:50)
Mun, I just had to ask— is there a specific example of Zafin innovation in corporate banking you could share?
Mun Tham (14:57)
Yeah, for sure. Lately we’ve been asked a lot about pricing optimization. A lot of bankers ask us: “Can you use AI to help us optimize pricing?”
The first thing we need to do—on top of centralizing product information—is centralize all the client relationships. Basically, we need to create the Client 360. So with that, we’re getting more into the lending space, bringing in lending pricing data into the platform to create this 360 client view.
For one particular client, we are looking at bringing deposit, liquidity, cash management, and lending data together to create the full view—so when the relationship manager is pricing a deal, they have the full picture, they know how much discount to give, and they can track the commitment afterward to ensure it remains profitable for the bank.
Cameron Smonk (15:52)
Great.
Cameron Smonk (16:02)
So I just wanted to get a feeling for how Microsoft and Zafin co-innovate differently—one obviously a huge technology company, one more boots-on-the-ground.
Lucy Li (16:16)
Yeah, I would love to comment on that. Think of Microsoft as coming in to help with the unified data platform and putting the AI layer on it. We build that infrastructure with the client—even in the trade finance example, we build the connector that allows banks to connect. Once we build a connector, any messaging a bank wants to send to the client can go out through that connector.
And what Zafin can do is push the pricing information so banks can communicate pricing to corporates. Microsoft focuses on creating the platform for the bank, and then Zafin comes in with specifics—what information should actually be traveling from the bank to clients or within the bank. You’re more application-specific, whereas we’re more platform-infrastructure-specific.
Mun Tham (17:24)
Yeah, yeah, yeah—I totally agree. Microsoft is at the infrastructure level. We help clients configure information onto that infrastructure platform layer. If I use an airport analogy: Microsoft builds the runway and the control tower, but we run the software to control flight schedules and ticket pricing. That’s how we work together to build this airport.
Lucy Li (17:50)
Yeah. For us, we look for patterns that apply to multiple workflows—less focus on a specific business application. For example, AI document processing can apply front office, middle office, or back office. But if a client comes to us and says, “We want to understand an application our bankers can use to manipulate or change pricing instantly—how will you do that?” we will refer the client to partners like Zafin.
Cameron Smonk (18:27)
Okay— and to end this off, I wanted to ask a rapid-fire, big-take question: what is your one bold prediction for the future of corporate banking?
Lucy Li (18:32)
I think in the near term, a lot of modernization—the hollowing out of the core—is going to take a very specific GenAI flavor. Banks should be looking for how the applications they’re replacing can embed AI capabilities already. That gives them a leapfrogging effect—they don’t have to build in increments. If you’re going to acquire a new platform, you might as well replace it with one that already gives you AI-embedded capabilities.
Cameron Smonk (19:34)
Great.
Mun Tham (19:35)
Building onto that: I definitely think data and AI are table stakes. Having data models, dashboards, copilots—those will become standard BAU. But the winners won’t be just the most digitized or the most AI-enabled. It’ll be the banks that can bring transparency to the economics layer, make it programmable, make it trustworthy for both people and machines—so humans can read it, but also AI agents can act on it.
Cameron Smonk (20:07)
Awesome. Well, thank you both so much for joining us at Banking Blueprints. My name is Cameron Smonk. Thanks for watching.
