Fast-Tracking Generative AI Transformation in Financial Services

Jim Marous (00:08):

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. As financial institutions face growing competition and increasing regulatory pressure, AI could unlock unprecedented efficiencies, innovation and growth, if only we can overcome the hesitation that holds us back.

Jim Marous (00:37):

In this episode of Banking Transformed, we sit down with Sandeep Mangaraj, Microsoft's Managing Director of Fintech to explore the critical role of speed in AI transformation for banking. Sandeep will discuss the barriers to AI adoption in the financial services, the risk of falling behind in this fast-moving technology landscape, and the tangible benefits of empowering employees with AI tools today.

Jim Marous (01:05):

He also shares insights into how banks can experiment with AI responsibly while building the foundation for long-term success.

Jim Marous (01:14):

AI has the potential to revolutionize banking and financial services, but many institutions still need to move faster to capitalize on its opportunities. In this episode, we'll discuss how financial institutions can avoid the pitfalls of hesitation and unleash the real potential of AI.

Jim Marous (01:33):

So, Sandeep, can you introduce yourself to our audience and share a bit about your background before we begin. Also, it'd be great to discuss a little bit about how Microsoft today is helping finance institutions navigate the landscape to improve the scale and speed of AI implementation.

Sandeep Mangaraj (01:49):

Excellent. Thank you, Jim, for having me on this podcast. I'm a huge fan of yours and I learn a lot. So, I'm very glad I'm here and I have the opportunity to share some of the insights from the job I do, from the work I do.

Sandeep Mangaraj (02:03):

I lead Microsoft's Fintech team, and we focus on financial technology firms of all sizes from startups to large established firms in helping them scale and grow with us. What that means, and what it has meant for the last two years is a lot of conversations around AI.

Sandeep Mangaraj (02:22):

So, a very pertinent topic indeed, at how we need to and take advantage of AI to be able to scale or grow our business, and this new world we are going to. Now, you asked a little about how does Microsoft approach it and what are we doing?

Sandeep Mangaraj (02:36):

Fundamentally, we are starting with trying to see what is it, what are the business outcomes that our customers are trying to solve. Because if we focus on what is it that the customers are trying to solve, then the tools, the technology, the people, processes follow, so we are leading with outcomes and a big focus on outcomes.

Sandeep Mangaraj (02:55):

Now, having said that, I'm proud to be working for a firm which has best in class platform, we have the full stack, and you can choose your own adventure. You can go as deep as you want, or you can take advantage of all the tools and technologies that we are bringing forth. It's a very exciting time to me.

Jim Marous (03:14):

It really is, and Microsoft really is on the cutting edge of this, and I recommend that our listeners, if you haven't done it already, go back over the other podcasts we've done with Microsoft. So, they've really covered a lot of different topics around how financial institutions can really take advantage of AI.

Jim Marous (03:32):

So, you wrote a recent article, Sandeep advocating for financial institutions and all enterprises to move from a crawl, walk, run approach which is a traditional way of implementing new technology to a run, walk, crawl strategy, especially for AI transformation. Can you explain a little bit about that and why is this shift so critical for financial institutions today?

Sandeep Mangaraj (03:57):

Yeah, I've been working, as I said, with firms of all sizes in terms of helping them address whatever is the pertinent need that they're trying to solve for. So, one of the things that we all need to have, make a decision on is what's happening with AI? Is this a fundamental platform shift or is it a continuation in the innovation that we have seen over a long period of time?

Sandeep Mangaraj (04:22):

Now, I personally believe that there's a fundamental platform shift, and this fundamental platform shift is coming from a few areas. The first one is that we can actually now interact with this digital world that all of us are increasingly dependent on using natural language. Now, that opens up doors for all sorts of people and we will, I'm sure during the course of this Jim, explore it a little more.

Sandeep Mangaraj (04:48):

The second is around data and AI and what's happening around data and AI. I have been in this industry for a long time. The first day I joined, the question was, "Hey, you know the data can become better. We have data silos,” fast forward 20 years, and we are still at the same place. And I can tell you, Jim, it’s not a technology problem, it’s actually a people-process problem, it’s actually a how do we handle change.

Sandeep Mangaraj (05:15):

The good thing about the technology capabilities today is that it's manyfold better than when I was in graduate school and I was trying to go through supermarket scanner data and my challenge was not whether I had enough data, it was, what can I do with all this data? We didn't have the tools that we're in a much different place there.

Sandeep Mangaraj (05:35):

And then the last thing that obviously is a huge driver of change today, is the way we are working, the way we are working has fundamentally changed. And that's something that sometimes lost in all the hype and conversations around AI and gen AI and what's going to happen. The way we are working is changing.

Sandeep Mangaraj (05:53):

And so, if you take those three into account and you agree that there is a fundamental platform share, the question is how do we approach it. And this is why, the reason I'm advocating for turning around the approach that we've used and why you start with running, walking, and crawling, is where we are today in 2024.

Sandeep Mangaraj (06:08):

There's a massive amount of investments that the Microsofts of the world have made in getting capabilities that you can use today. And so, the question is, if there is a fundamental paradigm shift, the way we work is changing, the way we are going to get insights is changing, should we be waiting for all this to fall in place and have and share a path forward that makes sense for us? Or should we be starting today with experimentation? Should we be focusing on trial and run?

Sandeep Mangaraj (06:36):

That's what I think is what you have to decide on. Should I be experimenting? Should I try to see what is it that I can do and what is that I cannot do?

Jim Marous (06:45):

It's interesting, Sandeep, you mentioned right at the beginning of this, and I agree so much. We talked about a lot on the podcast that the problem is not the technology, it's the people. And it sounds kind of rude, but the reality is, leadership's really got to decide whether or not they want to talk about AI or if they want to actually jump in and start playing with it. And especially in a regulated industry, there’s so many excuses for not moving forward. You can find them every day.

Jim Marous (07:13):

You can say, "Well, compliance isn't in place yet. Well, the competition isn't doing this or that,” all these different excuses for not moving forward. When you are working with financial institutions, what are the primary barriers that you see preventing banks and credit unions from accelerating their AI adoption and how can they overcome these challenges?

Sandeep Mangaraj (07:35):

A really good question and something we deal with all the time. So, Jim I want to make a distinction between AI, which financial institutions have been using all the time, it's not new, versus I think some of the challenges and complexities that come with gen AI.

Sandeep Mangaraj (07:56):

But if you look at what are the guiding principles that institutions have been using to adopt AI and use AI successfully, those guiding principles around having a risk-based approach, around making sure that there's transparency and accountability in making sure that it's fair and it's non-discriminatory and making sure that it's done in a safe and secure manner. Those guiding principles do not change when it comes to gen AI, and we know how to manage it.

Sandeep Mangaraj (08:21):

What does happen is that there are certain areas in which new risks are coming up and it's a question of making sure that A, we understand what those risks are tied back to the risk and control framework that has been so successful for us all these years and then see what, if any, changes need to be made.

Sandeep Mangaraj (08:41):

So, part of it is making sure that we get comfortable with what other risks that gen AI may introduce and then making sure that we are comfortable. So, let's pick on one topic because I find it really interesting, like hallucinations.

Sandeep Mangaraj (08:55):

The question is, is hallucination a feature or a bug. So, let's think about two scenarios, one in which I am talking … you Jim, you are an expert in banking, you're writing a new paper on a topic that you're interested in, and you are using gen AI to brainstorm.

Sandeep Mangaraj (09:19):

You could turn hallucination to a hundred percent and more and it actually then becomes a tool that helps you think outside the box, maybe get a new idea because you know what is good, you know what is what, you are an expert.

Sandeep Mangaraj (09:33):

Now their hallucination, I would argue, is actually not a bug, it's a feature. It helps you actually be more creative, be more productive, become more efficient. Now, that same hallucination, if it is giving a guide to a contact center employee to respond and someone who is new in career, then you need to make sure you have right cut guard rails. It's not going up and dreaming up things.

Sandeep Mangaraj (09:56):

So, again, everything has to be in the context of the use, the type of person that you're talking, who is using it, and then that show how your risk and control framework should drive.

Jim Marous (10:06):

That's very interesting, Sandeep because at the beginning … November of 2022 when gen AI really got introduced for the very first time, a lot of people said, it is just not working. It's giving me bad answers. It's rudimentary. And people, sooner or later in most cases, realize that it wasn't the answers that were the problem, it was the questions being asked.

Jim Marous (10:28):

I mean, I know in the way that I use it. I tell people that if you have writer's block today, you are not using generative AI. Because the reality is, anything you give to Gen AI will give you a response that can get your mind going.

Jim Marous (10:41):

And as you mentioned, you still need humans in the mix to be able to identify hallucinations and sometimes to use them to spur new thought patterns. I had a gentleman on the show about seven or eight months ago now that was writing a book with a lot of different contributors.

Jim Marous (11:00):

And instead of asking generative AI, “How can I make these chapters better?” he actually asked gen AI, “If you were to rate this chapter, each of these chapters with only one out of five stars, what was wrong with the chapter that you only gave it one star?”

Jim Marous (11:16):

It was a complete different way of looking at how you use the tool to make your thinking better and I think that's where we really have to look at it. We have to say, we still need the skill sets of each human to integrate with the power of gen AI to move us forward in leaps and bounds better than we have before.

Jim Marous (11:36):

And I think another challenge, and you mentioned it around the slowness of implementation, is we're fearful of what's being done. How are you seeing institutions when they're first starting their process, how are you seeing institutions using gen AI to generate better results?

Sandeep Mangaraj (11:57):

I'll get to that, Jim. I just wanted to pause a little on one of the points you made, which is around how your experience using gen AI and how the questions you ask actually has a big impact on the results you get.

Sandeep Mangaraj (12:08):

I believe ChatGPT is the best thing that happened and also the worst thing that happened to adoption of gen AI, especially in enterprises, because many of us stopped with that first experience, at models that were fairly nascent.

Sandeep Mangaraj (12:22):

Today, the world is changing every day. The underlying models have changed. There is a significant difference between GPT-3 and GPT-o1 that just came out. Now there's a significant difference and that's the kind of thing that sometimes has been lost in this deliberate approach on making sure that I am going to do this right.

Sandeep Mangaraj (12:44):

So, what happens is, and that's one of the reasons I've been advocating for this run, walk and crawl. By the time you get that perfect use case with the perfect metrics lined up and ready to go, the underlying models have changed and you're back to the drawing board.

Sandeep Mangaraj (13:00):

So, roundabout way of answering Jim, the way people have unlocked it is because they've gone and tried it. So, when I look at those who are ahead of the game today, they are not waiting for perfect, they are very, very big on experimentation.

Sandeep Mangaraj (13:15):

In fact, I have customers who come to me after a Microsoft announcement and even before our teams has had a chance to digest it, they're there. They're like, "I want access to it. I want to try it because I can see how it helps my product become better; my solution become better."

Jim Marous (13:33):

And it's interesting because it does save time and effort. And financial institutions, are you seeing them use gen AI to make better customer communications? Are you seeing them use it for better efficiencies, better back office, or simply better ideation for looking at innovation for products and services? What category are you seeing them using it the most for right now?

Sandeep Mangaraj (14:01):

Right now, if I look at those which have gone to production is in the first two, not in the ideation and creation, it's more in the first two. You talked about customer service, customer experience, that's been an area of focus. I'd say a lot of people actually started initially with employment, employee experience.

Sandeep Mangaraj (14:18):

Because they felt safer with when the human is in the loop in the construct that they're doing, that human is an employee versus a customer who may not have the same context and honestly the same tolerance for risk.

Sandeep Mangaraj (14:30):

So, employee experience has been a full area of focus, customer experience has been an area of focus and where we are seeing customers get a lot of value. In terms of ideation, creativity, it's been tested. I wouldn't say that it's not been tested, but that's probably not an area of focus if I were to look at those three areas.

Jim Marous (14:47):

So, one of the things financial institutions have to do is they're getting into using generative AI and more and one of the challenges they have is personnel. They don't have the people on team that really know the direction to go. And so, financial institutions are having to determine who's my partner going to be? Who am I going to partner with?

Jim Marous (15:06):

In a previous podcast, we talked to Ally, who has used Microsoft as a partner. But when you look at the implementation of gen AI solutions, how should organizations look at selecting a partner who will get them off stuck?

Sandeep Mangaraj (15:23):

So, for partners, when I think about partners, there are multiple partners that you can consider here. One is obviously infrastructure, provide you a technology platform and there, I think the fundamental questions you need to be asking is, do you trust them? You trust them with your data? Do you trust them with running your mission critical workloads? Are they reliable? Is it going to be resilient? Can I get access to it when I want it, when my customers need it? And then security is absolutely fundamentally important.

Sandeep Mangaraj (15:58):

So, I think for infrastructure providers, those are the questions you need to go through as you're thinking about who it is that I partner with. Now, those are table stakes. In addition to that, you'll have to obviously focus on the technology, like where they are, what does a roadmap look like? What does a future look like? And are they going to truly partner with us or is it going to be purely transactional?

Sandeep Mangaraj (16:19):

Because a partnership is a two-way street. Part of the partnership is that that infrastructure providers open to learning from you as much as they share tools and techniques for you to drive your outcomes. So, I think that's when it comes to infrastructure provider.

Sandeep Mangaraj (16:36):

Now you talked about skilling and skilling gaps. And then to bridge the skilling gaps, many organizations are obviously looking at advisory firms, consulting firms who help them. My personal opinion, the critical question for you is not only their knowledge, their domain knowledge, their ability to experiment, are they doing it themselves, so if that firm has really not embraced it, I think it's really difficult for you to work with that firm to transform your business.

Sandeep Mangaraj (17:04):

But in addition to that firm's ability to help you meet their needs, a key thing that you need to be thinking about is are they going to be helping you upskill better, learn better because I think that knowledge transfer is really important. And just like we use the partnership for an infrastructure provider, you need to have a two-way street, you need to have the same with your advisory partners.

Sandeep Mangaraj (17:28):

Now, the last thing is that you as a firm actually can upskill fairly quickly today because there's not a huge difference between the leaders and those who are starting today. This field is changing every day. So, as long as your people who are willing to learn and you provide a environment in which it's safe to fail, I think that's really, really important in organizations, you can upskill quicker than anything.

Jim Marous (17:52):

That's very interesting, Sandeep. Because we're seeing that the best organizations that we've seen using gen AI are between the biggest and some of the smallest. Because size really, if you pick the right partners, size really is an inhibitor because you can do so much and as you said, the leaders aren't that far ahead, and the followers aren't that far behind yet.

Jim Marous (18:16):

But change is happening so quickly that you really need to get up and running fairly quickly right now because the trains already left the station as an analogy goes. Sandeep, as you look in the financial service industry, what are some organizations that in your mind stand out as really doing some exciting things in the world of generative AI?

Sandeep Mangaraj (18:37):

Jim, thanks for this question. You know Microsoft mission is to enable every organization to do more. And I work with organizations of all size, so I'm not going to pick winners and losers, I won't name them here but I will tell you a little about what I see organizations that are leading where they are.

Sandeep Mangaraj (18:57):

One thing you have heard about test and learn, but I think just the test and learn environment is created where there is a tone from the top. The senior leadership team needs to be on board that this is a platform shift that is happening that is going to change the way value is delivered in my industry and in my organization. I think if you don't have that, it's really difficult for you to actually embrace this because there are a lot of uncertainties.

Sandeep Mangaraj (19:28):

The second is definitely an eagerness from people who are in charge, to learn by experimentation. We will not be able to get all the answers, we will need to embrace change, and we need to embrace some level of ambiguity and uncertainty.

Sandeep Mangaraj (19:47):

And I think the ones who are the most comfortable today doing it have always had a solid risk and control framework around the model and model governance. So, to them, yes, this is a new way, these are new kinds of models, but what we have in place, we can extend it, we can get comfortable around it.

Sandeep Mangaraj (20:07):

And then like what the Microsofts of the world are also helping in earlier some of the risks. So, some of the indemnification and things that we are doing that is also helping with adoption but nothing succeeds like actually seeing it wild, you got to see it live in the world, that is what will actually lead to ideas.

Jim Marous (20:29):

Well, certainly data privacy and security as you mentioned are significant concerns, especially in financial services. How can financial services organizations adopt AI while still mitigating risk, particularly in highly regulated environments like ours?

Sandeep Mangaraj (20:47):

Jim, I think I shared the statistics in one of the blogs that you talk about something like 78% of employees are what we talk about are bringing their own AI. We know something about risk and cyber risk and that is the weakest link can lead to massive, massive failures.

Sandeep Mangaraj (21:13):

So, I mean, 78% is a huge number. I think five people bringing their own AI in an uncontrolled environment exposes organizations to a lot of risk. So, that is where I don't think the approach to addressing risks is by shutting the door, because people see the value, they're going to experiment, they're going to try it. What we need to do is we need to provide sandboxes, soap boxes for them to be able to try it.

Sandeep Mangaraj (21:42):

And there you can control what is allowed, what kind of data you can use, what kind of controls you have. That's the better way to do it, in my opinion, which is to make sure that you provide safe and control environment for your employees to experiment.

Sandeep Mangaraj (22:02):

But not experiment in the theoretical sense of the word. It has to be experiment with real data to try to address real problems because otherwise, it just becomes a science experiment. I've seen a lot of hackathons, which do not lead to successful outcomes because people are not given the ability to solve the challenges that they deal with on a day-to-day basis. So, I think that’s what we need to do.

Sandeep Mangaraj (22:26):

We cannot just shut everything off, just give a bunch of synthetic data and people tell people, "Go figure out how you're going to use it." You put in the controls in place, you provide them safe sandboxes, and then make sure that they can solve problems that they are dealing with on a day-to-day basis.

Jim Marous (22:41):

It's interesting when we talk about generative AI at financial institutions, a lot of people talk about the ability to really make it so that there's the ability for employees to use AI tools to improve their lot, to be able to test and learn and have innovation. How do you allow employees to democratize AI, if you will, while at the same time making sure that they're not bringing unsanctioned AI tools into the marketplace.

Sandeep Mangaraj (23:12):

That's where I think the first — we talked about these three legs, so there are already available software as a service solutions. The copilots of the world, which you can use, and which are by design meant to address many of the risks.

Sandeep Mangaraj (23:28):

So, for example, if you use the M365 Copilot, the same controls that are currently available for IT administrators in terms of making sure that people have access to only the information that they can and they should have for their emails, for their SharePoint drives, it's the same that is inherited.

Sandeep Mangaraj (23:49):

So, what you are doing is that you are inheriting those controls, robust controls are already in place, and you can start experimenting. Because we have seen for example, so from telemetry, like people are using Copilot for M365, there 11% fewer emails.

Sandeep Mangaraj (24:05):

I mean, there's a lot of statistics Jim, I'm sure I can share, and we can put it into show notes about the studies that are coming up that shows the benefit you get. So, that's what I talk about the run phase. You can start today with all these copilots.

Sandeep Mangaraj (24:22):

The second phase is when you start going beyond readily available solutions to solutions that can be customized and you're able to customize it for your workflows, because the real value never comes from the data or insight, it comes from what you do 15 minutes before and 15 minutes after. So, that workflow is really, really important.

Sandeep Mangaraj (24:41):

And how you can use these tools, use these models in enhancing your workflows, that should be the next, and then the last, which is when you are ready. There are going to be fundamental shifts that are happening and will happen as new ways of doing work are going to be dreamt of by people and there are people who are working on it all the time.

Sandeep Mangaraj (25:04):

So, for that, you will have to then go and probably fine tune your models, build your own models, and so that's what I've actually meant by when I was talking about run, walking and crawling and the difference is that you don't have to wait for all — it's not sequential.

Sandeep Mangaraj (25:17):

You can do it all in parallel and it's who you are enabling your data scientists and people who are deep in AI, what they can do is very different from what you should be letting your frontline workers do and by the way, there's a lot of creativity in frontline workers, let's unleash that today.

Jim Marous (25:34):

Well, there's so many ways to use it and again, it’s a matter of balancing the risk and reward but I think as we look at the power of this, we also have to look at what's the return. A lot of organizations really don't know what they can measure, what they should measure, what the ROI should be.

Jim Marous (25:53):

And part of that is a lot of organizations aren't looking at where their destination is supposed to be. In other words, they're starting with a solution without looking at the problem they're trying to solve. So, when organizations are struggling to get to the ROI of an AI initiative, what are ways you can measure and demonstrate AI's value?

Sandeep Mangaraj (26:15):

So, I'll start with saying that ROI is too premature and probably the wrong metric to use today and let me share why I'm saying that, Jim. If you look at the return on investment, let's go down to how you calculate a return on investments. You need to have some sense of what the cash flows are going to be, what are the inflows and outflows.

Sandeep Mangaraj (26:35):

Today, the pace of change, pace of innovation, and the uncertainty in terms of where this is going, there's a lot of volatility in that measure. So, I think ROI can give you false precision and honestly, you will be missing out on great opportunities if you are focused on ROI at this stage today.

Sandeep Mangaraj (26:54):

Because the difference is that ROI is really, really good when I have a process that I've been running, and I want to make it more efficient. When the processes are going to change because the way we work is going to change, that metric is not going to be very helpful for you, that's why I say it's premature today.

Sandeep Mangaraj (27:11):

So, there are other ways in which we can look at it. One way, you can use real options. Financial theory are going to give you lots of different ways of measuring this and trying to quantify it. Personally, the way I like to think, the way we should be thinking about it as a portfolio approach.

Sandeep Mangaraj (27:32):

We have always had run the bank initiative and change the bank initiative, so we've always had dollars for change the bank. Now think about this change of bank dollars and where you are putting those investments and some of that investment should be in the running today some of them should be in the walking and some of them should be in the crawling.

Sandeep Mangaraj (27:50):

And that's when what you are doing is you're actually betting on your people. Because that's the other thing that I get passionate about, about, "Oh, AI is going to replace humans." We talked about Jim, and you mentioned that AI is about how we are approaching it is putting humans in the loop.

Sandeep Mangaraj (28:06):

Human in the loop is also not another bug, it's actually a feature because what AI tools are doing is they're actually adding to our ability to be creative, it's adding link to a ingenuity. You're betting on your people by enabling these tools to make them much better at doing what they want to do.

Jim Marous (28:27):

It's interesting because some of the ways that generative AI can be used is also with low code and no code platforms such as Microsoft's Copilot Studio. How can organizations best use this way of addressing AI adoption and create new solutions?

Sandeep Mangaraj (28:49):

Jim, it's interesting you bring this up because I was thinking about the interview you had with Charles Lamanna and Charles talked about Samit Saini who was in Heathrow. He used to work in security and then using our low-code platform, the transformation that he was able to drive in Heathrow, including him moving from his role in security to IT.

Sandeep Mangaraj (29:15):

Now, think about what Samit was able to do with what we had, not that far back, two years back to where we are today. The Samits of the world can now unleash their creativity by using natural language. Now, most local platforms, there's still some learning to be done. You need to be able to navigate through its interface, you need to be able to navigate through its menu.

Sandeep Mangaraj (29:42):

You need to know how you are going to be able to stitch together that process improvement that you're doing. That's good, that's good investment but think about when you can do that even easier because you have a natural language that you can use to actually initiate these flows and initiate this, that's where we are going.

Sandeep Mangaraj (30:00):

So, with Copilot Studios of the world and Copilot Studio is very specific, it's meant for building really good conversational agents using the power of AI in a controlled manner. There will be a million Samits, there are million Samits who can go and do it because all of us, know how to ideate, know how to talk, how to think through problems. And so, that's the promise of this technology. That the digital world becomes really, really accessible and democratized.

Sandeep Mangaraj (30:32):

And that's where Jim, I love the fact that you talked about you are seeing it at two extremes, the very large organization and the small organizations, this is democratizing for organizations too because through that same API call that costs the same, the biggest banks have access to the same tools that a startup has.

Jim Marous (30:51):

Especially when you have so many partners out there that are in financial services that allow you to take advantage of their experiences. It's like a GPS system, if you just tell the destination, don't have a GPS system, it's not going to tell you the best way to get there.

Jim Marous (31:08):

But if you find partners that are already working in financial services, you can leverage them to build new ideas, take advantage of opportunities that already have been tested and some really exciting things on the landscape going forward.

Jim Marous (31:23):

From your perspective, and you're really involved in this a lot more than I am, but certainly in Microsoft as well, what excites you about the future of generative AI in financial services in the near-term? I mean, we're moving so fast that near-term and long-term are sometimes mixed in together because everything's happening so fast but how do you see generative AI really changing banking in the near-term?

Sandeep Mangaraj (31:50):

So, there's a quote I love, it was by Justin Trudeau in 2018 at Davos. He said, "The pace of change has never been this fast, yet it'll never be this slow again."

Jim Marous (32:02):

By the way, it’s a saying I use in every one of my presentations, because if you think about it, it says so much about everything. It says so much about business, AI, politics, whatever it is, everything's moving so fast, you're right.

Sandeep Mangaraj (32:16):

In 2018, fast forward to 2024, it's not been truer. It's amazing and so this is where we have the opportunity to really rethink the way financial services firms do, what we all got into this business to do. I mean, it is really, really important for the economy that we have a robust financial services firm where there's competition, there's creativity.

Sandeep Mangaraj (32:43):

I mean, if you look at all the priorities that we've been focused on around payments, around accessibility, all of them, this technology can really, really change the way we bring it to market faster, better, more efficiently and better for the consumer. So, that's all great.

Sandeep Mangaraj (33:00):

The reason I started with that quote is that I wish I could predict where we are going because this is changing faster than even I have an ability to keep track of. However, I do know one thing, the ones who are going to be experimenting, who are already experimenting will be going far ahead of those who are taking a wait and watch approach here.

Jim Marous (33:25):

That's so key. I think it's a concern that I have as I look at generative AI and actually, all things having to do with change is that catching up gets harder and harder each day, not just for organizations, but people. I had a podcast I did with a friend of mine who we were talking about, a lot of people are being displaced right now for various different reasons.

Jim Marous (33:50):

And if a person wasn't prepared for the potential of being displaced, if they're not continually learning, they're falling further and further behind and putting themselves in a situation where human nature says you're going to start to blame others for your demise, for lack of a better word. And I think, and I will go to my grave saying that generative AI gives you more tools to learn than anything we've ever had access to.

Jim Marous (34:20):

It's like this Encyclopedia Britannica back in the day that used to have books and books and books that used to have to go through to find the simplest answer. Today, you have tools, multiple tools, there's so many different ways to look at it that can make your job easier, but the most important part is it helps you think better. It helps you learn better. It allows you to stay up with what’s going on and maybe even ahead of the curve, but it’s also taken away excuses for not moving forward.

Jim Marous (34:54):

If you are looking at financial institutions today, Sandeep, and they're just starting to get their feet wet, they're dipping a toe in the water, which is not your suggested way of doing things, what is the one suggestion that you give them today to make it, so they don't fall any further behind?

Sandeep Mangaraj (35:14):

The reason I pause is because, I've gone over many, you want to try now, experimentation is great. But honestly, it just comes down to embrace the uncomfortable, change is not always easy. I mean, Jim, I've gone through a few changes in my career.

Sandeep Mangaraj (35:34):

Financial services last 22 decades, we have gone from crisis to crisis, but we're stronger today. I really believe we're stronger today, and we are better off today. We have a more resilient system today. So, embrace, change, embrace what is uncomfortable, what's the worst that'll come out of it? You'll learn something.

Jim Marous (35:55):

Yep. Well, it's interesting because we talk about this often. I use this, it's somebody else's saying, but what doesn't kill you makes you stronger. The reality is, changes isn't going to stop, you're not all of a sudden going to get to a day, you go, "Thank goodness it's over." And we keep on referencing the fact that it’s only going to go faster.

Jim Marous (36:18):

So, while I was able to look at my career that in banking started in the mid-1970s, and I look at the amount of change between the mid-1970s and 1990s, it was moderate as I look back at it. I look at what’s happened.

Jim Marous (36:34):

I tell people today, if they went to university to learn about programming, if they went to university to learn about marketing, and you just graduated two years ago, the amount you don't know is almost greater than the amount you learned in that amount of time. That's scary.

Jim Marous (36:53):

On the other hand, and you mentioned it, if you've already experienced major change and you realize, "I'm still here. I'm okay. I may have liked to have not had some of these things happen to me, but I've learned from those," it is something you have to embrace.

Jim Marous (37:14):

You're not going to be able to avoid it and I think it creates a lot of anxiety. It creates a lot of stress; it creates actually what we're seeing in the world today from a tension standpoint because change, things are going to be different tomorrow than they were today. And we can’t go back, and we can only prepare for that.

Jim Marous (37:35):

And I sound like I'm preaching at times, but you've said so many of the words that I love saying, and I think it comes from what's happened in your business, what's happened in your career, what's happening at Microsoft.

Jim Marous (37:49):

I mean, I remember Microsoft at the very beginning of ChatGPT within the first, I'm going to guess it was three months, there was a major problem that happened that it started hallucinating in terrible ways. And as opposed to simply putting your heads in the sand and saying, "We're going to restart, we're going to start over again," you move from that in such a quick way.

Jim Marous (38:15):

And this was management that did this, move from it in such a quick way that most people on the podcast today will not even remember what I'm talking about. And the reality is, you have to adjust to the changes and the mistakes we make quickly, learn from them but don't stop moving forward because of them.

Jim Marous (38:38):

Nike came up with a saying, “Just do it,” but we can use it certainly in the application of generative AI that we just need to do it, test it, play with it cautiously, but not so cautiously to your point that we're crawling.

Jim Marous (38:55):

I suggest everybody on the podcast go to LinkedIn, look at the article that Sandeep did. It's posted on LinkedIn. It's not hard to read. It's not a hard read. It's good. It talks about the fundamentals of moving faster in the implementation of AI in regulated industries.

Jim Marous (39:15):

Sandeep, thank you so much for being on the show today, I really enjoyed having you as a guest. And again, I will do a seamless plug for previous podcasts we've done in conjunction with Microsoft, because every one of them brings another element of insight as to how technology and humans can come together for better solutions. Thank you so much, Sandeep.

[Music Playing]

Sandeep Mangaraj (39:37):

Thank you, Jim. It's been a real pleasure.

Jim Marous (39:40):

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 the show a five-star rating. Also, be sure to read my recent articles in The Financial Brand and check out the research we are doing for the Digital Banking Report.

Jim Marous (40:00):

This has been a production of Evergreen Podcasts, a special thank you to our senior producer, Leah Haslage and our audio engineer and video producer, Will Pritts. I'm your host, Jim Marous. Until next time, remember, change has never happened this fast, and it'll never happen this slowly again. You need to get on board and embrace the change that's before us.