What you'll learn
- How to capture higher quality deposit growth through digital marketplaces
- How to create consumer value through personalized, needs-based ecosystems
- How to drive digital leapfrog strategies for growth and efficiency through cloud and generative artificial intelligence (GenAI)
Ed Sander (00:10):
Okay, thanks everybody for joining. I'm sure the pizza was good. We've got a lively discussion to start with you. So EY is proud to be an innovator sponsor at this year's digital bank in 2024 forum, and we're proud to have our distinguished panelists up here to have really interesting conversation about capturing your right to win in this digital era. EY has done some research on this specific topic. Every year we do a pretty broad piece of research that we call our next wave study. I just want to share with you some bullet points out of that study that are pretty interesting for this conversation. We surveyed over 2000 retail consumer users, and what we found was that 48% of them are considering switching banks at any individual time. 64% actually of them when it comes to their deposit account, are also willing to switch their provider about, most of them have like four to five financial services products relationships.
(01:10):
56% of them actually do trust that their data is going to be used by a financial institution to curate the best experience and provide the best product recommendation for them. And when they actually are searching for information to choose a new financial product based on a life event, only 27% of the 2000 people that we surveyed already have the provider in mind that they want to use. So just some interesting headlines. We have that research study, it's available in your, you can access it through the app, but it's a good foundation for today's discussion. So let's start it off with deep. When you think about digital marketplaces as in an advent of digital ecosystems, what's changed pre and post pandemic? And why today in 2024, is it something that every financial institution should consider?
Dee Choubey (02:02):
Well, great question. So first of all, we started moneyline in 2013. So we've seen the arc of evolution over the last 11 years really accelerate in terms of the trust factor with data. So the priors had to be in place to lead up to 2024 where marketplaces actually can be useful. So if you think about what's happened over the last 11 years, at least in our case, we started off with really creating a direct to consumer digital bank. So this was a venture funded startup. FinTech, we raised over $300 million to really build out a customer acquisition platform that predicted when consumers would have money and when they would run out of money, and we built the entire digital bank around that. And over time, what we realized was that our secret sauce was actually more on the interface layer of actually helping consumers make the next choice for the next product, whether it's a personal loan, a heloc, a cd, a savings account, a mortgage, a credit card.
(02:59):
And as we built that technology and we invested hundreds of millions of dollars in building that technology, what we realized was that the core technology can now be embedded into anywhere the internet. So if you look at Moneyline today, we are a marketplace. We are a two-sided marketplace. On the one side, we have hundreds of publishers. So think about anywhere you consume your financial news, whether it's CNBC, whether it's Fortune or Forbes, you see that mortgage calculator or that personal loan widget or that which credit card is right for me. Oftentimes we are the technology's that's powering that. So we take that consumer lead, we cleanse that lead, we get all of the consumer consent, and we pass that along in a one to many ecosystem to hundreds of financial institutions. So if you think about that marketplace in terms of the trust that is needed, right?
(03:53):
So what's happened over the last 11 years, and again, the pandemic was a big C change in terms of really being a forcing function to increase the adaptation of this technology is that you have these APIs where more and more the industry is getting a common set of ways to communicate with each other. In the past bank A versus bank B versus FinTech A or FinTech C, there would be a lot of mistrust. I'm not going to trust you with the consumer's data. I don't want to lose the customer. But increasingly in this digital world, the priors are now getting in place for financial institutions to trust each other. And the regulators have actually, for the most part, we can gripe about things here and there, but the formation of how a company protects the data, shares it with somebody else, how we get the consent from the consumer, all those priors are now getting in place where you can really have true marketplaces, right?
(04:53):
Because financial services still at the end of the day is like a flea market, right? You get clicked from link to link to link, and each forum you're filling out not getting the benefit and the consumer experience, the consumers filling out the same information five times. But now in these digital marketplaces, the consumer proposition is really clear. And for banks, we can now by embedding these marketplaces, really keep your LTVs higher with your consumer. Don't lose those consumers to the money center bank. If you're a bank that's only in a certain asset class, but your consumers are looking for a credit card, you don't want to lose them to B of A, right? So you would embed a lot of these digital marketplaces inside of your own ecosystem with your rules. And that's really the evolution of the business of moneyline has been that we're as much a compliance and a regulatory RegTech piece of technology that sits in the middle of that network. Then we are really the customer acquisition engine. So we now generate 80 million leads. How do you get access to those in a way that's compliant in your own ecosystem? That's really the evolution now, and I think we will talk more about this later in the conversation with all the advancements in artificial intelligence. It's a really interesting place now where all of the rules of how you share this data is come into play
Ed Sander (06:13):
With gen AI. It's something that we will talk a little bit about. I know there's got some thoughts about that, but
Dee Choubey (06:19):
I didn't want to steal his thunder.
Ed Sander (06:21):
When you think about digital marketplaces as something that's new and novel, we talk a lot about the new and novel. What do you see in the market from money lines, conversations with different financial institutions? How are they thinking about digital marketplaces and what would you share with the group today?
Dee Choubey (06:37):
Well, look, I think that initially there's skepticism, right? Because I think if you look at a lot of the largest financial institutions, they're closed loops. We will start with Chase, right? They have their own mortgage, they have their own credit card, they have their own personal loan, they have their own insurance solution, they have their own. So they don't want the consumer to leave the ecosystem. But the rest of us, the 4,000 banks, mid-sized regional financial institutions, we may not always do everything ourselves. We started off as a digital bank. We really kind of deemphasized the lending part because we didn't want to go and raise wholesale and securitization capital, but we still wanted to have a relationship with the consumer. And I think that's what these digital marketplaces really enable. They extend your customer's lifetime value. You can now generate non-interest income through your existing client base through commission revenues.
(07:24):
You can now influence their next best offer. They can come back to your bank first to make decisioning, right? You can provide them a much more trustworthy experience in terms of KYC onboarding. And of course, if you look at what's happening on the policy side, 10 33, rule 10 33 really now codifies that the data, it belongs to the consumer, they can passport it from whether it's a FinTech platform, whether it's bank A, bank B, bank C, right? And I think the larger institutions have fought it a little bit, but I think the cat's out of the bag in terms of whose data it is, it's the consumer's data. And I think that is going to actually make it much more consumer friendly. And you see GDPR, you see a lot of the privacy regs. Those are all required to get us into a place where institutions and consumers and marketplaces can share that data in a copacetic manner that the regulators also signing off on.
Ed Sander (08:21):
So interesting, just this culmination of new technologies in the digital efforts by institutions. Yesterday there was a session actually about one institution, S-E-C-C-U, I think that has opened up a virtual branch as a way to service its customers. So Antonio, you own Gen AI at Citizens Bank. Just tell us, what is Citizens Bank been doing? What have you been brought in to do? Where are the current focuses today?
Antonio Iñiguez (08:46):
Yeah, I was brought in end of November, 2023 to lead the generative AI team. So really I think the approach from citizens is to take a very responsible approach to this new technology. So the first thing that they did was stand up a centralized leadership team that consisted of a cross organizational group of individuals from our chief experience officer to our chief technology officer, chief risk officer, to help us identify, prioritize generative AI use cases, and then also plan how we were going to deliver on those use cases in a safe and responsible way.
Ed Sander (09:20):
Can you just expound a little bit more on those use cases? Are they, how are you operationalizing them?
Antonio Iñiguez (09:25):
Yeah, absolutely. So originally I think we took a list of around 90, and Michael Rutledge just recently came out with an article that talked about our process. But one of the big things that we wanted to focus on for our first foray into generative AI was does it fit within our risk appetite, which currently our risk appetite is we have to have a citizens colleague in front of the generative ai. And then also, is it going to be a high value driver, whether it's efficiency, productivity of our colleagues internally mixed with that. We also wanted to find a use case or a set of use cases that were extensible. So we build it once for a very targeted use case, but then can expand to other areas. I think Satish yesterday did a great job talking about their summarization use case, which can be re-leverage probably across their organization and a number of other summarization approaches.
(10:12):
And we took the same approach. So we started with two use cases. The first one was for our contact center as well. Our contact center agents, probably some of the hardest working individuals in the organization. They have to consistently answer calls from our customers and deliver a high level of customer service. And one of the challenges that they have, and one of the friction points that they have in their role is identifying and finding information to answer our customer questions. So the first use case that we have is our knowledge management application, which is a natural language question and answer system that given a question from our contact center agents, it'll quickly scan our policy documents and provide them with the best possible response as well as references. So a very common use case across the financial industry from what we found, but it's already had rave reviews and our colleagues are really excited about how that's going to change the way they work and in turn impact how we're delivering customer service with removing the friction from our colleagues, having to look up this information will enable them to then provide that higher level of service to our customers.
(11:15):
So really excited about that use case. Our second use case is something a little bit more internal. It's a leveraging GitHub copilot from Microsoft to empower our developers to work on more challenging problems. I think that's one of the key things our leaders really want us to focus on with generative ai, is freeing up our colleagues time to work on more impactful, interesting problems to then provide them with a better experience in their role, and then tangentially impacting our customers as well. So with coist is the use case that we're talking about here for copilot, our developers, and there's probably some developers in the room. I'm a former data scientist. Before I started this role, I can attest I did not like writing unit tests and I did not like writing documentation. So that along with finding those answers that I used to search Stack Overflow for is one of the key things that we're trying to do for our developers. So those are the first two use cases that we're delivering internally. We're in pilot currently, so our contact center agents, we have a small select group of folks who are answering customer calls, leveraging the contact center solution. And our developers have also had access to COAs Assist for a month or so now, and they're using that on their day-to-day.
Ed Sander (12:25):
Yesterday, our keynote speaker talked about how the entirety of their institution is really signed up to some of the AI initiatives. When you think just about your colleagues at Citizens, what's the overall mindset towards bringing AI in as a backbone for operations?
Antonio Iñiguez (12:44):
I think there's a ton of excitement, I think, across the organization, and I think it goes back to what I mentioned with our leadership is really emphasizing that generative AI isn't replacing people's jobs. It's providing them an opportunity to do their job differently and taking some of the friction points away. So if we look at our developer team, we're not going to necessarily remove developers from our organization, but they could do things more quickly, work on more challenging problems and take away some of the monotony of their day-to-day writing unit tests and documentation to enable them to work on those more challenging problems. And I think that messaging that we're sending across the organization through our Talent and change department has really resonated with our colleagues. There's a ton of buzz around citizen around generative ai, and people are knocking at the door to get their use case in our prioritization list.
Ed Sander (13:32):
That's really inspiring to hear. Maybe just one more thing. From an institution's perspective, the topics of trust and integrity and controls around the use and the deployment of Gen AI seems like it's always a sub thread in every AI conversation. How is that sitting inside of citizens? What's the dynamic of those conversations?
Antonio Iñiguez (13:54):
Yeah, and again, this goes back to I think how our leadership team approach generative AI is a program and not as isolated pockets of individuals in the organization delivering on generative ai. So I mentioned that centralized program that we have, it includes our risk compliance legal partners, our model risk validation teams. So safety and trust are at the forefront of every decision we make, and it also, we consistently reevaluate our risk appetite as we get more used to this new tool. It's an exciting piece of emerging technology, but we're taking a responsible approach while also trying to innovate quickly. So I think just having them at the table has been immense in enabling us to find that safety and trust while leveraging this new technology.
Ed Sander (14:34):
Okay. Maybe one last question for you, Antonio. So Dee talked a lot about how digital marketplaces and money lines journey of bringing that service into the market to touch the decision capabilities that people have at their fingertips is really being impactful. When you think about retail consumers and what AI can offer their experience, what do you see when you're kind of looking ahead?
Antonio Iñiguez (15:00):
Yeah, absolutely. And I think what we're going to see a lot in the near term is impacts to the consumer. Not necessarily giving them access to generative AI via chatbot or anything like that, but backend processes that are enhanced that will then enable the customer and consumers to have a better experience with our organization. If we take a look at the first use case that I talked about, contact center knowledge retrieval use case, it's really going to enable our contact center agents to take away that friction point and provide a higher level of customer service. I think Satish mentioned this yesterday, or in the video they mentioned this, where their contact center agents are also freed up from having to remember every single piece of the conversation to write that summary. So they're able to deliver a better experience for our consumers. And I think that's where we're going to see a lot of the generative AI use cases in the near term. As far as the long term, this is a quickly evolving technology, so we don't know where it's going to go at this point. It'd be kind of throwing just darts at a map and hoping we land on the right spot. But I think overall, we're going to see a ton of enhancements to the customer experience leveraging generative AI by enhancing those backend processes, taking away friction points and delivering better customer service.
Ed Sander (16:11):
Thanks so much, Antonio, there. When it comes to Microsoft, Microsoft has an army of darts to throw at the dashboard map, what is Microsoft's view for gen AI and how it's going to be a disruptor in a positive, constructive way just in the market at large? What should institutions think about in the audience, kind of learn from Microsoft's view? Yeah,
Uzair Hussain (16:38):
Sure. So I mean, I'll take a couple angles to this, and I think some of the themes that you've heard over the last day and a half will resonate. If we think about just broadly how gen AI is reshaping financial services. We can talk about things in three horizons. So I think we've spent a lot of time on horizon. Horizon one, your classic in the four walls of the bank, human in the loop productivity use cases, and you reflect on the current macroeconomic environment, cost optimization pressures for banks, exacerbated by nim erosion, competition for deposits, higher loan loss provisions. It all makes sense because you're scratching inclined to get dollars in through the door and get positive operating leverage out of those dollars. So those productivity use cases make a ton of sense in the contact center call summarization, sentiment analysis, the whole chat with your data use case to get knowledge much more quickly.
(17:30):
We're also seeing a lot of uptick with customer facing personas, so commercial bankers, investment bankers, wealth advisors, the ability to more efficiently generate a pitch book, a customer presentation, a financial plan is certainly having a lot of legs. If we shift to horizon two, and Antonio, you touched upon this a little bit, and we can loosely call this putting a generative AI chatbot directly in front of a customer. I think naturally we are seeing's more of an incident infancy. We're seeing some hesitation from banks, which makes a ton of sense just given the nascency of the technology itself. Certainly chatbots have existed for quite a long time, but they're fairly limited, right? If a customer zigs instead of zagging, you have a human in the loop to have to solve that problem. But with generative ai, you'll be able to have a much more natural language conversation with your customer.
(18:20):
And I see that happening in a couple of different ways. The first would be being able to chat with data on a customer or on a bank's website, right? Tell me more about your deposit products. Compare credit card A to credit card B, so I can make an informed decision in kind of the awareness, awareness path. However, we are seeing a couple of banks do POCs around using a gen AI chat bot to actually execute on a servicing transaction, whether that's blocking a lost or stolen credit card or helping to execute on a credit limit increase, as an example. The third horizon, which is probably the furthest out and a bit more difficult to predict, is I believe at least generative AI will actually introduce a new class of financial products, a new type of servicing model, new value propositions. And as we are co-creating with our customers to build their roadmaps, I feel confident that those use cases will start to manifest themselves
Ed Sander (19:15):
At EY a lot of what we do in the market is helping our clients when they're undergoing some type of transformative initiative. From our conversation over dinner last night, I know that you do a lot of travel helping and facilitating fis when they're thinking about transformation and adopting Gen ai. What are some of the observations that you would share with the audience?
Uzair Hussain (19:33):
Yeah, sure. So I mean, I'll start with a little bit of a horn toot if that's okay. I mean, so generative AI mean it's been around, it's a relatively recent phenomenon, but if you think about AI more broadly has been around for the better part of 30 to 40 years, and Microsoft has been investing in AI for the better part of 30 to 40 years, which is reflected in the 22,000 AI patents. We have. One of the reasons why OpenAI decided to build their models on our infrastructures, just given our large data footprints, you've really got the horsepower to make these models models run. But to answer your question directly, and I appreciate you letting me toot the horn a little bit, I think we're thinking about it in a few different different ways. So we are building a copilot experience, an AI powered assistant for every role and persona in the bank.
(20:26):
And it really starts with the copilot for publicly available data. And the difference between that and some of the free versions that exist out in the market is your data stays in your four walls. Anything that you search does not go out into the ether or train some of the foundational models. Then you have the copilot experience into our productivity suite. So hopefully in the products that you use every day, teams outlook, PowerPoint, Excel, word, and then being able to reason across all of your chats, all of your emails, all of your documents, all of your files to really drive that productivity, then you can extend that experience to some of the partners that you work with. So being able to chat with your SAP data, your Adobe data, your ServiceNow data, your Salesforce data, and the last piece, which I think is really the holy grail, is being able to chat and deliver insight from your proprietary bank data, your transaction systems, your product systems, your systems of record, and of course doing all that in the most comprehensive, secure and compliant way, grounded in our set of AI principles. So that's how we're seeing from a Microsoft perspective, generative AI really deliver value across the bank.
Ed Sander (21:43):
Building on that Uzair, if there's one aspect or area of investment that an FI would be contemplating, because as Antonio shared, there are many different use cases that an institution can choose from, but we're thinking about making a difference, providing innovation that distinguishes that bank. What are some of your observations from your conversations in the market?
Uzair Hussain (22:06):
Yeah, fantastic. One of the things I'll highlight is Microsoft recently commissioned a study called our 2024 Work Trend Index, where we survey about 31,000 people across 30 different countries globally. These are leaders for a number of different organizations, and two key data points that came out loud and clear. One 79% of those surveyed felt that generative AI has the potential to create sustainable long-term advantage in their organization. However, less than 50% actually knew how to measure that or how to actually get started, which is a fairly large chasm to navigate. So if I think about a three-pronged approach to executing or implementing on generative ai, I would say thing one is empower your employees with some of the shrink wrapped first party out of the box AI assistance that already exist in the products that you use, and do that in, of course, a safe and controlled environment.
(23:06):
Because what that does is it starts to build the muscle, it starts to build the new ways, new ways of working. It generates power users that can share learnings with other folks across the organization. And all of this, of course, as citizens did, is governed by a centralized cross-functional AI council. So that's sort of step one, I think step two is once your employees get a lot more comfortable with AI in those new ways of working, how do you then take your underlying LLM, add on a low-code solution platform to then extend that or customize that to your workflows, to your business objectives, to your pieces of data, which will then be the next tranche of value? In my perspective, those two categories will account for at least 80% of the massive backlog of generative AI use cases that many of you I'm sure already have.
(24:00):
Now, if you think about sort of thing three, and to your question around differentiation, there may be some scenarios where it is worthwhile to actually invest in fine tuning a model or actually building your own model. And while that is extremely resource intensive, it does provide the highest potential for differentiation for long-term sustainable competitive advantage to be able to build a proprietary solution to differentiate from the competition. But ultimately, I think that approach in the short term, what it does is it reduces the risk of shadow AI because we know from all the folks that we've researched, employees are bringing their own AI tools to bear. And that can be an issue from a risk and compliance perspective. It also helps just build a culture of AI and an operating model to help govern that. And of course, it also helps then be able to track and measure and realize the value of your overall AI journey.
Antonio Iñiguez (24:57):
And if I could add on, I think you said something really, really important, getting the champions is something that we found is absolutely essential when we're rolling out our generative AI use cases. So providing an opportunity for the colleagues or fellow employees to test out these tools to get them comfortable leveraging this, whether it's on their personal devices at home, not leveraging your company data, just to get them used to it and start to think about it as a performance enabling tool as opposed to something that's going to replace them. So getting them early access and enabling them to be champions for you is absolutely essential, as we've learned at Citizens in our first two initial use case rollouts.
Uzair Hussain (25:36):
Yeah, great point. The change has to be both bottom up and top down. Absolutely.
Ed Sander (25:40):
On this point of distinguishing, DI know from our conversations leading up to today, I mean, digital marketplace is certainly democratized choice, right? For the masses. One of the themes that we've talked about that you've expressed some interest in that's important from market perspective is this concept of financial wellness and personalized financial management. What are your views about what institutions are doing there to also provide that differentiated service?
Dee Choubey (26:10):
Yeah, I think that a lot of the things that Uzair was talking about in terms of the tooling and the instrumentation of all these AI models, we're actually now bringing that to bear in the application layer. So we always say that we're a technology company, we're a data company. A lot of the work that Uzair is saying is hard. We're doing that for both the consumer, but we're doing it on a white label basis that anything that happens on the money line front end, you can download the app and you can play around with it. It is the most advanced talk to your money application that's out there, right? So whether you want to talk to your own transactions now with 10 33, we're now able to get your brokerage account data, your bank account data, your credit card account data, your mortgage data, and just the example that you were saying was there, right now I can have a full conversation about my spending across my entire life as opposed to just my transactional data with my money line bank account, or my Citizens bank account or what have you.
(27:10):
And that's really powerful because now with your credit score, with your full asset and liability picture, the PFM tools, we're going to leave mint.com in the dark ages here very soon. And I think we're already there. And I think the elegance of what we are building at moneyline is we're not making it a closed loop ecosystem. So the hard work that's required with making sure that the data lineage is correct, that it's copacetic from our compliance and regulatory manner, our APIs, even our low code capabilities now are going to be made available through our partnership with EY as well as some of the other things that we're doing to any bank financial institution to leverage that, whether it's content. We've invested a lot in financial literacy. A lot of people talk about it. We actually have a studio where we produce even the one oh ones of investing, banking, mortgages, PFM. We now make that available as a feed to any financial institution. In fact, we're doing some great work with even MasterCard where we're powering their digital feed now. That's going to be made available to everybody. So a lot of the hard work that Microsoft and open AI and others have done on the backend, we're doing the hard work to make that very easy to ingest for both consumers, but also on a B2B basis, any bank or financial institution that wants to use it.
Ed Sander (28:30):
We're certainly seeing that a lot in our client conversations in the market as well. Institutions are shifting with laser focus, their investments in any core modernization effort to really have distinguishing investments in the digital channel. Okay. Panelists, for this packed room, we're going to close out our conversation. Please offer your headline view. What is your prediction on the future of gen AI, digital marketplaces, digital channels? What would you impart to the audience of something that they will see in the headlines in the months ahead and who, and if we talked a lot about this last night, let's start with you.
Uzair Hussain (29:09):
Yeah, sure. I'm happy. I'm happy to start. I would say generative AI is certainly a transformational capability, but again, it's a tool in service of a business problem or a customer problem. I think right now, folks see it as a hammer and they just see nails everywhere, and they want to use it for every type of issue that exists, keep the business problem and the customer problem top of mind. And there may be an opportunity for generative AI to solve that. There's a whole host of other tech and tools that can also solve very similar problems. So that's what I would leave you with.
Ed Sander (29:40):
Ed Antonio, what's your view based on the journey that citizen has taken so far? What would you share with the audience today?
Antonio Iñiguez (29:46):
Yeah, I'll concur with what ER said. Honestly, it's all about gen AI plus other technologies. So I think what we'll see is there's going to be a lot of low hanging fruit that that industry has gone after, currently, the summarizations contact center knowledge management use cases. But what we'll find is that the truly transformative opportunities will required generative AI plus some other technology, whether it's OCR, robotic process automation, the up and coming multi-agent approach for generative ai. It's all going to work together to really drive transformative opportunities where generative AI will play a part but not be the sole driver of that change.
Ed Sander (30:19):
And Dee, you're out on the market often. What's your view?
Dee Choubey (30:22):
Look, I think it's the age of cooperation, right? I think that partnering with companies that are here, a lot of the gap that exists in the tech spend that the money center banks have an advantage on, and they're just able to spend billions of dollars a year in doing it themselves. We're enabling everybody in the room, every bank, every small midsize financial institution to catch up and take advantage of the hard work that we have done on cleaning the data, making it compliant and usable by both the consumer and on the back end. So it's the age of cooperation, I think.
Ed Sander (30:55):
Okay. Well, on behalf of ey, I'd like to thank all of you for joining us through this afternoon conversation. It was well worth the wait after the pizza. And I'd like to thank Dee Uzair and Antonio for joining us. EY was proud to be a sponsor at American Banker this year, and we look forward to coming back again and see you next year. Thank you everyone.