Bridging the Generative AI Readiness Gap

Jim Marous (00:07):

Hello, and welcome to Banking Transformed, the top podcast in retail banking. I'm your host, Jim Marous, founder and CEO of the Digital Banking Report and co-publisher of The Financial Brand.

Jim Marous (00:18):

AI and generative AI are certainly set to revolutionize the financial services industry. But how are banks leveraging these powerful tools and what challenges do they face in the implementation process?

Jim Marous (00:33):

Monica Hovsepian and John Radko from OpenText will share their insights on the current state of AI and banking, its potential to transform customer experiences, and strategies for fast and seamless deployment.

Jim Marous (00:47):

Drawing from the latest Digital Banking Report, they'll explore the current adoption levels of AI and banking, the challenges institutions face in implementation, and the strategies for bridging the AI readiness gap. I guarantee this is a session you don't want to miss.

Jim Marous (01:07):

AI has a potential to revolutionize banking and financial services, but many financial institutions need to move faster to capitalize this opportunity. In addition, they have to prioritize where their deployments are going to take place.

Jim Marous (01:24):

In this episode, we'll be discussing how financial institutions can get the most from generative AI solutions and where they should start.

Jim Marous (01:31):

So, Monica and John, before we start, can you introduce yourselves to our audience and share a bit about your backgrounds? Monica, why don't we start with you?

Monica Hovsepian (01:41):

Hello, and thank you very much for this opportunity. I'm Monica Hovsepian. I work for OpenText, and we were the sponsors of this excellent paper by Jim Marous. I am the head of financial services marketing at OpenText. I've been with the organization for nearly six years now, and I'm based out of Toronto, Canada. With that, John Radko.

John Radko (02:03):

Thanks, Monica. My name's John Radko. I'm the senior vice president of engineering for OpenText. I lead the global software development team that builds our content suite, our experience suite, our business network, and our platform products.

John Radko (02:17):

I joined OpenText through acquisition and I have about 30 years of industry experience. So, really excited.

Jim Marous (02:25):

So, as Monica mentioned, we just completed Digital Banking Report on actually the trends and the status of generative AI in retail banking. And what was interesting, the report showed us that 84% of financial institutions believe AI will provide significant benefits, yet only 12% have a well-defined roadmap to getting to where they want to go.

Jim Marous (02:50):

So, Monica, what do you think is causing this disconnect?

Monica Hovsepian (02:56):

It's looking at this and saying, yes, there's a lot of room, but not everybody is really ready to start wrapping their arms around it, and they don't know where to start to go with this.

Monica Hovsepian (03:07):

Now, one of the things that you actually shared in the report — now, I don't want to give everything away as to what's in the paper, but there was a report by Citibank that you actually shared in the report, is that 93% of financial services institutions expect that AI is going to improve their profits over the next five years.

Monica Hovsepian (03:31):

And if now, we're going to translate this into numbers, is that AI can boost bank profits by 9% or 170 billion by the year 2028. Now, it's not all just around generative AI, but it's how AI can actually help drive potential of profitability.

Monica Hovsepian (03:55):

Now, in your report, you also, quoted that a staggering 2/3 of all the work done in banking and insurers has the high potential for AI driven automation and augmentation. And this is higher than any other industry.

Monica Hovsepian (04:12):

And if we go back to, for example, 2016, and we take a look at InsurTechs that came into the market such as Lemonade and Clearcover, for example, they have been coded to say that at least 60% of their work has already been digitized. And they were already using AI before they came into the marketplace. So, there's a lot of lessons learned right off the bat there.

Jim Marous (04:43):

So, John, from the same perspective, looking at there's this gap between where people want to be and where they really are. Everyone says there's a potential to change the world, but how does OpenText see this happening and how is your organization helping to support this change to get people more ready for the implementation?

John Radko (05:06):

So, I think one of the challenges is people tend to focus on the technology and they think a lot about LLMs and machine learning and all that. And where they really need to start is at the use cases.

John Radko (05:17):

And I think the most exciting opportunity automation is going to bring lots of productivity, and we've seen that applied in fraud detection and in other places.

John Radko (05:25):

But I actually think the real opportunity is about service at scale, at bringing that kind of personalized level of service that typically only the top clients or the biggest clients receive and being able to deliver a similar experience to everybody. And the way that begins is by identifying the specific use cases.

John Radko (05:47):

So, a great place to start is with our approach to marketing. Instead of blasting the world with sort of plain vanilla offers, can we take all the information we've got about customers and tailor offers specifically to them, to the point where instead of being like an intrusion or spam, this is actually welcome because it's a perfect fit for the customer.

John Radko (06:11):

The way we try to help with that, it's really twofold. With our content aviator, we help customers engage with information they've already got, whether it's documents or structured data. So, think of this as a salesperson or a customer service agent who can actually ask questions of their own content.

John Radko (06:30):

And this is a relatively out of the box solution where it's able to look at existing documents and look at that. So, that's leveraging existing information.

John Radko (06:40):

But beyond that we're looking at systems that integrate sometimes programmatically through APIs with other information sources to overcome some of the big barriers around legacy integration and whatnot to pull all that information together.

John Radko (06:56):

So, that's our enablement stack and technology. But the key thing really, and I just can't hammer this enough, is start with the use cases and the value you're trying to deliver to that customer. And really think about in the retail cases, where can we deliver personalized service at scale?

Jim Marous (07:14):

It's interesting, John, because you reference, not only do we need generative AI to tell us where we are, but the really important element is how do we have the rubber hit the road? How do we really deploy this? And OpenText, I know, has developed a lot of solutions to help actually deploy it so the customer feels value.

Jim Marous (07:34):

I talk about the old days when I was in banking actually as a banker, and I'd have to work so hard to get data, but it was really data to see where we were as opposed to data that could be deployed immediately for the benefit of the consumer. And I think this is really where the holy grail is. It's in the deployment, not in knowing where we are.

Jim Marous (07:54):

So, Monica and John, the report shows that almost all organizations believe generative AI will improve customer experiences. Monica, what do you think is driving the focus on using AI to improve customer experiences? And I'll move it over to John as well.

Monica Hovsepian (08:11):

One thing that you said, I think it was last year or the year before, is that we're all looking for that GPS of financial services. You actually did a video for us. And that message has truly resonated with me. And that's exactly what we're looking for, especially now.

Monica Hovsepian (08:27):

We're all looking for our financial services organizations to provide that financial health and financial wealth for us. We want our financial services organizations to provide us with that guidance, understand me, know who I am above and beyond that KYC.

Monica Hovsepian (08:46):

Know who I am and provide me with that guidance at the right time, in the right place that is contextualized, personalized, and is relevant to me. Don't bombard me with offers that are not relevant. And if I've already said no to you in one channel, don't offer it to me again. My bank keeps on doing that to me all the time.

Monica Hovsepian (09:13):

So, take that into consideration. And it takes a lot more than five or six or seven variables to understand who the customer really is. And now, we're putting so much information out into so many different channels, utilize that information and direct that information specifically to me.

Monica Hovsepian (09:37):

And that's exactly what we're looking for, especially in this economy and how things are progressing. Now, we've got a newer generation. We've got the millennials and we've got the Gen Z, and they are digital first.

Monica Hovsepian (09:49):

Utilize all that information that they're also, putting out there to become relevant. Because if you don't, they are now going to go to that neobanks and that digital first banks that are truly recognizing the value of that information and how to utilize it. There is other industries that are truly taking the value of that information and targeting that customer base.

Monica Hovsepian (10:15):

Now, you actually referenced the Bain study in our paper that said that customers are now, starting — they don't see that they're getting that relevant personalized content back.

Monica Hovsepian (10:28):

J.D. Powers two years ago said that only 44% of retail banking customers are getting that personalized service. The latest report from Bain is saying it's actually a lot less than that. And that is from a global survey.

Monica Hovsepian (10:43):

A lot of us are recognizing that, but we are getting personalized content from the likes of Amazon, for example, from the big techs, from Netflix, and we're getting it from Uber. They know exactly where we are, what we're doing, and they're targeting us. So, why aren't our financial services organizations? That's all we're expecting.

Jim Marous (11:06):

John, Monica brings up a good point that the consumer's getting more and more used to getting a very high level of personalized experiences. But even more than that, they're getting new levels of engagement.

Jim Marous (11:20):

Uber's a great example Monica brings up because I went to Amsterdam a little bit more than a year ago. And as I got in the car, Uber, they know where I was going for my location.

Jim Marous (11:33):

They said, “Would you like dinner to be delivered when you get there? And here's some restaurants we could have it delivered from, and it'll be there when you get there.” And I thought that was really a good link in to what they do from the food side.

Jim Marous (11:46):

Well, on top of that, when I didn't respond, they said, “Well, if you're about to think about going out to dinner, here's some places we recommend.” And they were very targeted to the type of food that I eat because of their relationship with OpenTable and being able to see what I want. And then they're already booking my travel back.

Jim Marous (12:02):

And it's very interesting how consumers almost subconsciously are getting used to the levels of experiences. John, from your perspective, what you see in the marketplace, what can banks do better in leveraging AI tools, leveraging data, leveraging the advanced technology we have at our fingertips right now, to build better personalized experiences?

John Radko (12:25):

Well, that's a great question, and where you need to start is to understand the bank's primary responsibility is that customer relationship. That's where it begins. That's the best place to differentiate.

John Radko (12:40):

Today, financial institutions largely have very similar capabilities, very similar scale, but the way they interact with customers makes the difference. So, understanding what's our brand? How do we relate to the customer?

John Radko (12:55):

Borrowing your Uber example for a minute, Uber's about customer convenience. If you think about every example you shared there, it was about delivering convenience to you. So, what is the bank trying to do?

John Radko (13:08):

And then it becomes a race, quite honestly. How quickly can I get this capability out versus my rivals? And part of that is I want to devote 80 to 90% of my focus and intelligence into figuring out how to better serve the customer, how to bring my capabilities in. And then I want to force the vendor and my technology partners to bring me the capabilities that do that.

John Radko (13:35):

So, for instance, rather than say deploying a general AI tool from a frontier company like an OpenAI or a Google, and then trying to integrate it to all my data. No, no, make your technology partners do that for you.

John Radko (13:49):

So, in our case, we do a lot of content management. Make your content management company bring you that integrated capability, and then you spend your time tailoring that to the use cases you've defined.

John Radko (14:01):

So, it begins with defining what do I want to do for the customer? And I think Monica hit it. There's a lot. It's personalized wealth management, it's advice, it's offers that are tailored to my life situation there.

John Radko (14:17):

We have so much data. So, the focus is really around integrating that data, bringing it together, demanding the utmost of our technology partners and understanding first and foremost, what is the brand, what is the customer relationship message we're trying to deliver?

Jim Marous (14:34):

John, that is so key. I was talking to somebody yesterday, I said, this industry of ours over the last five, seven years, it's amazing what technology partners, solution providers have been able to do on behalf of their financial institutions.

Jim Marous (14:51):

In fact, the research we've done and the people we talk to, we find that some of the most progressive financial institutions today are some of the smallest. The reason being, they have leadership that says we need to change the way we do things, and then they go out there and they find the solution providers that can get them up and running quickly.

Jim Marous (15:09):

We were having discussion as well with a person from Australia yesterday, I was saying in my day to implement even the simplest of changes took a year or two years, sometimes three years. We can now, implement the most advanced changes on a narrow area of what we want to do in banking in three months, in less than three months.

Jim Marous (15:31):

It's amazing what the capabilities are. And as you said, picking your partners that you can work with, such as OpenText that can get you up and running the quickest, helping to avoid the pitfalls that they've seen, you've seen with partners that haven't done it maybe quite right.

Jim Marous (15:47):

And taking advantages of the speed of deployment that you've seen and the successes you've seen. So, we can get that lowest hanging fruit faster than we've ever been able to get before because of the capabilities of working with partners.

Jim Marous (16:02):

So, I'm going to shift a little bit. Another part of the report showed that 68% of respondent banks were still in the beginning stages of AI journey, and nearly a quarter of the banks have not even started yet. So, in my world, I say a lot of talk, sometimes a lack of action. And I think there's a lot of reasons.

Jim Marous (16:24):

But, John, why do you think banks are facing the slow adoption of AI?

John Radko (16:31):

Well, I think there's a lot of (I'm going to use a big word here) fear. And I think there's fear about breaching trust with customers. I think there is a fear of how expensive it's going to be. I think there's a fear of the high cost to get started and the integration challenges.

John Radko (16:56):

So, I think a lot of it is fear of how to roll this technology out in a way that's safe and that won't degrade the brand. So, I think that's one of the biggest factors out there.

Jim Marous (17:09):

I think you're right. Monica, from your perspective, what have you seen to be why is it so hard to move from great dreams to actual implementations?

Monica Hovsepian (17:21):

I think that John really touched on it. I keep on saying this, the most powerful five letter word in financial services is trust. And I think every single financial services company needs to truly concentrate on the word trust. Every single person that works in a bank or insurance company is responsible for building trust.

Monica Hovsepian (17:45):

So, when it comes down to AI, there is this fear of does our customer trust AI? And then how do we implement it in the right way to then develop and build on that trust?

Monica Hovsepian (18:00):

Having said that, just yesterday, CNN did an article around the US Treasury has now implemented AI. Yes, it was machine learning, but as we're talking around AI, it seems to be this all-encompassing word until you start breaking it down.

Monica Hovsepian (18:18):

They did implement machine learning to check fraud in just fiscal 2024 of over $1 billion. Just wrap your mind around that. And that's over three times what they had done in the year before.

Monica Hovsepian (18:39):

So, if we take a look at the value add of how AI can actually be used to assist the public, that's already a lot of value add.

Monica Hovsepian (18:53):

Now, they did also, say that it wasn't just AI that they were utilizing, there was always a human involved in the mix. So, it's increasing the potential of the human to then deliver trust and assist. I think that is where the value comes in.

Monica Hovsepian (19:13):

So, we've got to take a look at incredible use cases of how we can augment the human potential to deliver better outcomes for customers and to build on that trust. And that just one use case.

Jim Marous (19:28):

Yeah, it's great, Monica. You both brought up the whole issue of trust, but it really also, gets to the point of saying the consumer's going to give you more information, more insight into their needs, their desires as we deploy solutions that respond to those desires.

Jim Marous (19:45):

I ask people in conferences I deliver at that, “How many of you have Amazon Prime?” And virtually everybody raises their hand. I said, “It's amazing you're spending 130, $150 a year for Amazon Prime, but I bet most of you won't even be able to describe what you get from that, what value you receive over and above.” It used to be free delivery, but they get free delivery for almost anything right now.

Jim Marous (20:13):

And the reality is, it's really trust. Amazon has not broken our trust by losing our data, having bad actors out there. They continually make it so that my selection process is easier.

Jim Marous (20:26):

And I think it's the fear of missing out. People are afraid, myself included, to stop that relationship for fear of what I may not get, because I really like my relationship with Amazon.

Jim Marous (20:38):

So, John, we've referenced it, we've danced around a little bit, but security is always top of mind. And we've just seen recent news about banks and bad actors and things that have happened negatively using AI as well.

Jim Marous (20:51):

How do you see that banks can possibly build a higher level of customer trust around the sharing of information, but also, our use of AI to make their lives easier?

John Radko (21:04):

There are a number of approaches banks can take to that. The first is to explain to customers what they're doing, to be as transparent as possible. So, if you're leveraging an AI capability as part of a service you're delivering, be open about that.

John Radko (21:19):

In many cases, customers expect it. And if it's done correctly, I think it's a positive experience. And help customers understand the benefit of it.

John Radko (21:28):

The second one is take security seriously. And this is interesting, we talked about banks being conservative, but I actually also, think financial institutions are among the best when it comes to data security.

John Radko (21:41):

And granted, there are a lot of news stories about breaches. Let's keep in mind people try to breach those institutions because of the data they have. It's not because they're sloppy. Of our customers, I would say banks are uniformly the best at information security and they need to bring that over.

John Radko (22:01):

Something else, though, when I talked about demanding from vendor partners, AI features that are integrated into existing systems are going to, in general, be more secure than AI that it is bolted on from the outside. And the reason is that they share the security model.

John Radko (22:21):

I'm going to give us as an example, we are by no means the only one, but in our content aviator, which allows people to summarize documents or generate new data from documents that is integrated with our document permissioning.

John Radko (22:34):

So, if someone at a bank is doing a summary to look at what a customer's business is with the bank, that summary will be based solely on data that individual is allowed to see. The bank has existing security protocols in place and they're able to leverage them.

John Radko (22:50):

The number one thing to watch out for when you do an AI project is that you don't accidentally subvert your own security. And the best way to do that is have AI features that are already integrated with that security model. So, those are some of the ways.

John Radko (23:06):

There's a lot of things that go into it, but one is to be very open with customers about what you're doing. So, be transparent. And the other thing is to honor that trust by leveraging all the strong security protocols you have in place.

John Radko (23:21):

Oh, one other is to monitor and audit. This is a time period when you have to keep a really close eye on things and make it easy for people, both internal and external to give you feedback because that feedback can enable you to react quickly when something unexpected happens.

Jim Marous (23:39):

John, I love the way you started that conversation that said, be transparent. You also, brought it back at the end of your comment.

Jim Marous (23:45):

And I think the consumer knows we're using their data. The consumer knows we're going to be using AI. The consumer knows there can be mistakes made. But if you're upfront with them as opposed to making it so they don't really know how their data's being used, the problem's going to be when something goes wrong and they're going to say, “You know what? I trusted you with this.”

Jim Marous (24:07):

And I think we can build trust by being open, by being completely transparent and sharing with them what we're doing on their behalf.

Jim Marous (24:15):

And I come back to what we discussed with regards to trust earlier, is that if you show them value of exchange, if you show them better ways of working with them, better ways of solving their problems proactively, as Monica said, without selling, but by recommending based on what we know about them, eventually they're just going to get really comfortable with the fact that you're working on their behalf and you're trying to make their life easier.

Jim Marous (24:43):

As Monica mentioned, my reference to a GPS of financial services, it's the difference between simply showing me how to get from point A to point B, to using crowdsourcing to say, “There's a traffic problem ahead, here's an alternative route you may want to take, or here's your shortcut, here's an avoidance.” Those elements really improve the trust in the system to get it right more often than not.

Jim Marous (25:07):

So, it's interesting, in the report we found out that 69% of financial institutions prefer to buy and deploy third party solutions versus building them internally or even partnering directly with the tech provider.

Jim Marous (25:24):

From both your perspective if you're going to put on your banker's hat right now, what factors would you consider when choosing a partner between these approaches? Monica, I'm going to start with you on that.

Monica Hovsepian (25:38):

Obviously, has the technology, a partner who's a proven vendor, a partner who's actually done it in the industry. And if I were the buyer, a partner that I know has the historical background that I can partner with throughout the entire lifecycle and is willing to assist me in this journey and that can provide me with that guidance.

Monica Hovsepian (26:08):

That's what I'm looking for. Insights and somebody that we can hand in hand take this journey together. Again, it builds trust. It's a relationship.

Jim Marous (26:21):

Yeah. John, from your perspective, what do you think we need to do to really — what would you look at, I should say, when you're selecting partners?

John Radko (26:31):

So, there's a few factors. One is, do I have a track record with the partner already? Have I worked with them? Have they come through? How deep is their bench? How deep is their experience? Do they understand my industry and my use cases, and what's their specialty?

John Radko (26:50):

So, in other words, we consider ourselves to be the information management company and we have a lot of experience in financial services. So, when we're working with our financial customers, that's what they expect from us, that we're going to help them manage their information, and that we're also going to understand their problems and be able to speak their language.

John Radko (27:08):

Another key one is to understand their roadmap and their history. So, where are they going and where have they been? So, now, that obviously favors existing companies and there's definitely a place for the bright-eyed startup. But when you're talking about integration or service at scale, you need someone that's proven their ability to scale.

John Radko (27:33):

The other thing is how many other customers do they have in my industry? And the reason that's important is one of the advantages you get from a vendor is that you're benefiting from the experience they're having on all their customers.

John Radko (27:47):

It's not that as a vendor we're better than our customers. Many times, our customers have amazing IT staffs. It's just that we gain experience by working with hundreds and thousands of customers, whereas an IT team is only working with one customer by definition. So, it's really that scale, the trust, the record, and their vision. Where are they going?

Jim Marous (28:10):

It's amazing, as I mentioned before, the depth of knowledge, the skill sets, the ability to help institutions of all sizes in their journey is something we've never seen before in the industry. What's nice about it is it allows you to deal with composable solutions in a specific area.

Jim Marous (28:30):

So, you can build small, you can build big, but you can partner with an organization that's already gone through all the trials and tribulations of deployment of an AI solution such as OpenText.

Jim Marous (28:41):

So, now, we used to end these conversations with what do you see in 5 to 10 years? Well, nobody has a clue. I mean, the reality is change is happening so fast and it'll never happen this slowly again. So, we got to shorten that timeframe.

Jim Marous (28:58):

So, I'm going to ask each of you, how do you envision AI transforming banking the next three to five years? So, Monica, why don't we start with you.

Monica Hovsepian (29:08):

I think we're going to see AI being utilized to assist the humans a lot more. We're going to see a lot more digitization. I do not see the human being replaced. I am seeing that the human interactions are going to take over for more complex, more advisory roles.

Monica Hovsepian (29:27):

And I'm seeing a lot more personalized engagements driven by AI. I think we're going to now, start seeing embracing AI a lot more as we develop this trust with AI.

Monica Hovsepian (29:44):

We're seeing it in other industries in different ways as we're wrapping our arms around, again, more generative AI, we're embracing it more. I think we're going to see the industry embrace it more as well as we pivot.

Jim Marous (29:57):

What do you see, John?

John Radko (29:59):

So, it's really in two areas. I see a growing focus on service at scale and productivity at speed. And let me explain this. I think service at scale is what we started talking about, which is that ability to bring that sort of private banking service level to as many customers as possible.

John Radko (30:18):

And I see it where banks increasingly compete on their customer experience. And it's going to be a combination of amazing automation because customers expect the same degree of digitization and automated tools that they have with people like Amazon, as you said earlier, with their banks, but also, that personal touch.

John Radko (30:39):

And what's going to enable that personal touch is what I mean by productivity at speed, which is the ability to help the human beings that compose that financial institution come up to speed on an area, or a customer, or a sector very quickly using tools.

John Radko (30:57):

And I think what we'll see is the competitive advantage of a financial institution will be disproportionately linked to their ability to use these tools in the service of their strategy. So, it's not generically using the tools, but how well can they wield them to deliver the service level that they aim to deliver to their customers.

Jim Marous (31:20):

Monica, John, I so appreciate you being on the show today, and I especially appreciate the sponsorship of our new report, The State of Generative AI in Banking that has been released.

Jim Marous (31:32):

We also, are going to link the link to the report within our host notes. So, if you go back to the way you entered this discussion, you'll be able to hit a quick link to get access to the report. Again, Monica and John, thank you so much for your time and I really think we'll be discussing this again.

Monica Hovsepian (31:50):

Thank you.

John Radko (31:51):

Thanks, Jim.

[Music Playing]

Jim Marous (31:53):

Thanks for listening to Banking Transformed, the winner of three international awards for podcast excellence. If you enjoyed today's interview, please take some time to give our show a five-star rating.

Jim Marous (32:03):

Also, be sure to catch my recent articles on The Financial Brand and check out the research we're doing for the Digital Banking Report.

Jim Marous (32:12):

This has been a production of Evergreen Podcasts. A special thank you to our senior producer, Leah Haslage; audio engineer, Chris Fafalios; and video producer, Will Pritts.

Jim Marous (32:21):

I'm your host, Jim Marous. Until next time, remember, now more than ever, financial institutions must understand the power of AI and take immediate action to meet market needs.