Fighting AI Fraud with AI: A Three-Pillar Playbook from United Asia Finance’s CIO CIO Leadership Live: ASEAN · 24 May 2026 Source: https://chrt.fm/track/3B598B/traffic.megaphone.fm/FCI8292514994.mp3 Transcript: https://podcasts.opennash.com/shows/cio-leadership-live-asean/fighting-ai-fraud-with-ai-a-three-pillar-playbook-from-united-asia-finances-cio-25f2f7/ (via OpenNash Podcasts, machine-generated) Welcome to CIO Leadership Live, ASEAN. In a city where a single deep-fake video call once cost the company $25 million in a single afternoon, the pressure on financial services technology leaders has never been more consequential. Hong Kong's consumer finance sector sits at a fascinating crossroads. A digital native challenges, raising the bar on customer experience on one side, and increasingly sophisticated AI-powered frosters on the other side. Caught in between are millions of consumers whose trust and financial well-being depend on the decisions made by the lenders they rely on. My special guest today is Jim Munn, Chief Information Officer of United Asia Finance. He has been building something remarkable. In just two years, with a dedicated IT team of 100, Jim has delivered three transformative innovations, an AI document defender that cuts loan verification from hours to seconds, Hong Kong's first generative AI chatbot in financial services, and a real-time data lake powering smarter decisions across the entire business. Jim, welcome. It's a great pleasure hosting you today. Yeah, it's good to have a conversation today. Yeah. Today, Jim and I will dive deep into the story behind his strategy, his candid views on the challenges shaping the future of financial technology across the region. Jim, could you please start by sharing with our CIO audience your background and career trajectory? How did you rise to become the CIO at UA Finance? Yeah, I started my banking career 20 plus years ago in UBS and has been on a journey with UBS and front office international. And then I lead the digital transformation for JP Morgan for decades in Asia and then move on to retail banks and build up virtual bank in Hong Kong and also move on to be our lending business right now and has been in UA Finance for four years as CIO. So I see different waves in terms of technology and banking industry going on and together with the help of and and now with the agenda ai era um and previously there's of digital transformation and nlp coming in and then now the genvi um and data lake data uh integration is one of the core pillars in the banking in the online industry so and now on uh we use digital use AI to help customers in terms of servicing, in terms of provide a better offering to the customers and also maximize the value curation in both the industry, the banks or the money lender and to the customers. So in old days, he was talking about using phone calls, using a lot of paperwork, but now with technology, it's moved everything with the bank and the lending. Thank you, Gene. Jim, deep-fake fraud incidents in Hong Kong's financial sector has surged dramatically in recent years, culminating in the shocking US $25 million incident in early 2024. From your vantage point as a CIO in consumer finance, how do you see this arms race between AI-powered fraud and AI-powered defense evolving over the next one or two years? Do you think the industry as a whole is moving fast enough? Yes, it is. Yes, we need, it's not too fast, but it's fast enough. We need to catch up more on those latest technology. You know, there's a, in different space, the technology in terms of fraudsters or cyber criminals is huge. The pace is very, very fast. We need to catch up on that kind of defense. we can't be just using the old one that will be not using the white shield to defend the sword. Completely agree with you, Jim. Now across Hong Kong and the broader Southeast Asia region, the shortage of fintech-savvy talent is consistently the top barrier to digital transformation. As the CIO leading a team driving some of the most advanced AI initiatives in consumer finance, How are you approaching the challenge of building and retaining the right technology talent? And what do you think the industry is getting wrong? In my perspective, I will definitely use those agentic AI to help us, help our developers to do those vibe coding. We have got a lot of AI to do those automating for those. In old days, there were days to build a skeleton of an application. And today we just use a lot of prompts as a genetic solution to just ask the agent to create their code, create the automation and check their code, do the merge request, have certain gap wells in all these quality control. And then we better use this technology to amplify our capability of our tech team instead of finding lots of developers to do tedious work. We are just mastering the user requirement, mastering what is good look like in both technical and non-technical, those non-functional requirements. We have to cater all this and ask the agent to build this stuff. It will be much faster and more complete and maintain a good level of quality in terms of deliverables. This is a key point. Yes, true. It's so difficult to find this kind of talent in the industry. It's a way in the industry right now. we need to use more technology to attract this talent instead of telling them all using the old ways of doing use the great decode in old ways you have to do all the coding repetitive way not using AI so we invest in using AI to help us to do those as I said, white coding in cybersecurity protection and data analysis to send back. Thank you, Jim. Now, as I know, Hong Kong's government only released its policy statement on responsible AI in financial markets in October 2024, meaning that many institutions, including those that deploy Gen AI checkbots and fraud detection systems, were operating ahead of the regulatory framework. How should CIOs in financial services be thinking about responsible AI governance when the rule book is still being written? and where do you think the accountability sits with the regulator, the institution or you as a CIO personally? Yeah, talking about accountability, yes, definitely as a responsible company, we definitely take the responsibility in terms of how we do the business by using AI. In our case, we have built AI governance framework here one year ago and this framework will cover not just technology, cybersecurity perspective, but also those ethical, make sure those AI is fair to our customers, fair to our employees. And there's other considerations like the personal data, whether you can use personal data to ingestion the AI. That will be another topic to be discussed. And also those considerations regarding compliance, we also cover, because from time to time we have to comply with those internal and external policy in AI. And we have a working group, not just IT, but also business users, SME, to review all these AI use cases from time to time. You know, the easiest answer is not using this, but we have to use AI to amplify our capability. So we have to make a risk assessment on this adoption of large language model, for example, or in-house developed model. There's a comparison and proper risk assessment on this model in AI governance. We also refer those, as you said, there will be a lot of new benchmarking in the industry. So we have to reference and compare notes from time to time. So there will not be a one-off assessment. there will be continuous assessment in this AI governance model. And once there are certain risks exceeding, the company's risk appetite, that will be bring up and discuss it in detail. So this is the way that we have to be very cautious to using this AI, but we will use it. One of HKMA's own FinTech 2030 findings is that integrating new AI solutions with legacy infrastructure remains one of the most resource-intensive barriers to innovation across Hong Kong's financial sector. For CIOs at established financial institutions like yourself, how do you balance the pressure to innovate at speed against operational reality of legacy systems and technical debt? Yeah, sure. Not every legacy system is compatible with AI. In my case, there are some legacy systems components that is not really scalable and safe to interact with the AI. So we make better assessment. For example, some program is single thread. Okay, if we do it in single thread, if not AI, it will be very slow. So we have to make sure this application is scalable, multi-thread, and also have a certain cybersecurity protections and when AI is calling these services. But we have to modernize different types of legacy components to be cloud-native, API-friendly, become a friend of components to interact with the AI. We become an MCP server for AI to understand if it's a tool for the AI to call. Do you think there's a risk that traditional lenders fall behind structurally? Sorry? traditional lenders is there a risk they fall behind and not be able to catch up I can say we wanted them to catch up as an equity committee leader I want everyone to be a better use of technology to catch up on these things yes it's a trend to use technology to help customers because it's what customer expecting and company to help them no matter if it's a lending or banks or retail stores what customer expectation is different You know, you just use a super app in everyday lives. It used to be digital solution helped you. So what a customer expectation in terms of lending process, they expect it to be very smooth on the app or on the web. And if we are talking about other play in the industry, definitely they have to be catch up in this kind of services. Otherwise, it won't survive in this different types of industry. stream. Agree with you, Jim. Now, real-time data platforms are increasingly the foundation of modern financial services innovation. Yet, close to 90 % of financial institutions globally cite data management, fragmented sources, quality issues, and regulatory constraints as their biggest AI adoption hurdle. How should financial services CIOs reframe how their organizations think about data, not just as a compliance obligation, but as a genuine strategic asset? Yes, you talk about the real-time data. It's another game changer in data integration. In other words, we're talking about data warehouse that takes end of day, end minus one data. It will not be very helpful to make instant decision in the fast-paced lending industry. So the live database has to interact with all those applications and it will become very slow. With real-time data lake, we can just do all those real-time dashboard. Those marketing automation just rely on this data lake. It can easily find the customer preference. It can give personalized triggers to the customer. Say if your customer wants get in the stores, not in the stores, in a convenience store, we place order and then immediately all those in-depth increase in real-time and back-end, and then we can do all those marketing automation to service up this customer to buy more and you go to the next store in that particular mall. So this is all this real-time distancing happening with the help of real-time data lake. Yeah. Thank you, Jim. Now, Hong Kong's digital banking landscape has grown rapidly, with virtual banks and fintech lenders raising the bar on customer experience and speed. Yet, recent surveys now show only around one in three Hong Kong users consider a digital bank their primary account. Do you see the rise of digital native challenges as an existential threat to established consumer finance players or an opportunity? I can't say spread. I think I say there'll be a shift in customer expectation and behavior in banks. It's already one third of customers in digital banks. In Hong Kong, I think there'll be even more right now in terms of this kind of one third spread. You know, the government also use a lot of roll out a lot of initiative that encourage hong kong people to use digital um in different schemes i can't say there's a competition if every additional benefit there's a advantage in in terms of their cyber security on or protection controls um the brand and and digital bands have a strong agility to roll out a new product the new band from time to time have a higher agility to roll out this new product and have testing tasting new user interface experience gamifications and as it can't say there's a competition there's a there may be a shift for customer expectation on on on digital um they're not competing in in the same same game Thank you, Jim. Now, I gathered from your CIO 100 Awards 2025 submission that you led three interconnected innovations. The AI fraud document defender, Hong Kong's first generative AI chatbot, and a real-time data lake, all delivered within two years with a team of 100. That is an ambitious portfolio by any standard for me. Can you walk us through the thinking behind this three-pillar approach? Yes. Number one, the document defendant, you know, we have been doing lending. We do a lot of customer diligence. And you also mentioned about, previously we talked about the more it fostered in the industry. So having this kind of frog defendant help our staff to check their, the customer submission, document submission and more carefully with the help of AI, And rather than all this in human, by human eyes, there's a human error prone, they may be missing in terms of verifying these documents. And with AI, there will be more consistent way to identify fraud documents and help the operator, which is efficiency of our staff, minimizing the risk of having a fraudster in becoming our customers. That's number one. Number two, the use of real-time data lake, as I mentioned. We are doing data business. Because, you know, we are underwriting a loan. We need to have all the data, the data business. Having a data lake to mash all this data together, not just making a bit underwriting decision, not just one or a few data points, but a very large data set of particular customers and this segment to make sure when we give a loan to customers or we can have a right changes rate and the customer can really repay their loan. So this is a very important foundation for our business to continue to provide more accurate decision, accurate product offer to our particular customers. The third one about the JANF AI chatbot. Yeah, we have been facing a lot of customers inquiry on the phone calls or a branch network asking what's the latest offer in our business. And, you know, we've generally produced all those RAG to understand our existing product offering and can answer the customer this kind of general inquiry, the latest offer inquiry immediately. So from time to time, we don't have to train our chatbot with manually placing those rules. the board can understand our product document and then answer the customer inquiry in a more accurate manner. For example, our offer is a time-limited offer. And if a customer is asking an offer, we won't send them an already expired campaign. So, you know, in ending, we have a lot of promotion all year round on different types of promote seasonal offer. And with AI, we can answer the question, oh, yes, this is this kind of Chinese New Year offer is already expired. And now we have Easter promotions. So this is very important. Otherwise, we have to put a lot of effort to write all those code, write also rules. And with the help of AI, we can all this automation and understanding answer customer inquiry automatically and also helped our, as I said, there will be a lot of agents answering calls at this kind of latest offer, save all these calls with our customers. And our agents will be doing more, even more value, add the services to our customers. Thank you, Jim. Now, for it to become Hong Kong's first generative AI chatbot in financial services sector is a remarkable first mover achievement. But first, our rally is straightforward. Can you share with us what were the hardest problems you had to solve technically, operationally or organisationally? One of the key points is I want the board to focus on answering only our products. We have to pay certain guardrails, avoiding the board to answer different types of something that you don't want. Just focus on helping a customer in UA Finance product and instead of answering something that is not on behalf and no capacity to answer those inquiries. So we need to configure what should the board should answer and what should not do it in any case. Now, you also highlighted measurable outcomes. Fraud attempts markedly decreased, call center workloads reduced and customer satisfaction improved. How do you as a CIO make the case for technology investments to your board and stakeholders, particularly for innovations where the ROI is harder to quantify upfront, like your Gen AI chatbot or the real-time data leak? Yeah, we do assess the ROI, no matter if it's an AI or other solution. return not not necessary uh money okay there will be brand awareness and there's also operational efficiency not and and also in our case cyber security also use an ai so this is um if you threaten all this as a as a return okay there's not just money okay put all these factor into our to return on the ROI calculations. The investment, yes, we need to calculate all these costs because, you know, the token, the crack token in the AI tokens, everything you have to spend it. You have to keep trading, monitoring, otherwise it will be going into chaos if it don't limit what you need to spend. If you should have a budget on these AI tokens and exceeding, you have to have a reaction plan on on what should what's to do next otherwise if you don't want a checkbook certainly don't answer your question because of uh on the use of all the um token credit so that would be a plan b in case this go out of control so yeah go back to your question how are we do calculate um return not just business the money but also other factor to come in for the cost yes we concluded but have to be careful factoring all those credit plans and the old infrastructure costs. Jim, thank you so much for such an enjoyable and insightful conversation. You're welcome. What emerges from our conversation is a portrait of a CIO who articulates the most powerful innovations are the ones that earn trust quietly, reliably, and at scale day after day. Jim's three pillar strategy, fraud defense, customer intelligence and real-time data is not a collection of projects. It is a deliberately interconnected system, each innovation reinforcing the other, all anchored firmly on the business outcomes. To our CIO audience, this has been Estelle Quagg together with Jean Munn, the CIO of United Asia Finance. Thank you. Thank you.