OpenNashPodcasts

AI in Financial Services

Why Banks Need a Chief AI Officer - With Ian Wilson

1 Jul 2019 · 23m · Daniel Faggella, Ian Wilson · 22 quoted facts

▶ Play episodeOriginal source ↗

In brief

Ian Wilson, a former head of AI at HSBC, argues that a bank should not have a single AI strategy. Each business line, such as retail, commercial, private and investment banking, needs its own AI strategy that supports that line's business goals. He says a group-wide leader, which he calls a chief AI officer, should set direction. The business units, not IT or data scientists, should drive the work.

The central tension is between early chaos and disciplined progress. Banks first spent money across scattered AI groups and are now asked where the return is and what comes next. Wilson says the fix is to start with dull, low-risk improvements to existing work, such as reading paper documents. As the business gains knowledge and comfort, it can then reimagine its operations. The host adds that most AI vendor funding today goes to risk.

Key insights

  1. Banks need AI strategies per business line

    Wilson says operations, retail, commercial, private and investment banking each have different strategies, so each needs its own AI plan that feeds its business strategy. One bank-wide AI strategy would miss those differences.

  2. Scattered early AI spending produced chaos, not ROI

    Wilson says banks threw money at many separate AI groups with no structure, so executives now ask what comes next and where the return is. He compares it to how mobile banking spending played out.

  3. Phase one is doing today's work better

    Wilson describes two phases: first do what the bank does today but better, then reimagine the business. He expects business teams, not AI specialists, to do the reimagining once they have seen what the tools can do.

  4. Paper-reading automation is boring but high-return

    Wilson says much banking work is reading paper and typing it into databases, so automated document reading is a low-risk start. He says a million documents read in an hour instead of a month could open up new uses.

  5. Business leaders must drive AI, not IT

    Wilson says business units often hold the budget and know what they need, so they should pull in capabilities rather than have technology groups push them. He adds that comfort is a hurdle, and seeing competitors use a tool removes the doubt.

Questions

Why does a bank need more than one AI strategy?

Ian Wilson says each business line, such as retail, commercial, private banking and operations, has a different business strategy. AI should support each one, so a single bank-wide AI strategy is the wrong approach.

What is a chief AI officer and why would a bank want one?

Wilson describes it as a group-wide role that decides what each business line should do with AI and how to carry it out. He says it brings structure to what was scattered spending, though the job need not carry that exact title.

Where should a bank start with AI?

Wilson recommends low-hanging fruit that improves existing work, such as using vision and language software to read paper documents instead of typing them in by hand. He calls it boring but a very large return on investment.

Who should lead AI projects inside a bank: IT or the business?

Wilson says business leaders should lead because they know what they need and often hold the budget. Host Dan Faggella agrees and says initiatives should not be driven by IT or purely by data scientists.

What are the phases of AI adoption in banking?

Faggella says his research sees three phases over about six to eight years. The current phase is risk reduction (fraud, compliance, cybersecurity), and efficiency comes in phase two.

Facts

Stated by the speakers. Each quote matches the transcript, which shows what was said, not that it is true.

Emerj's AI in Industry podcast had run for more than four years by this episode, interviewing AI experts across sectors.

“AI in industry has been around now for four plus years where we interview AI experts in various sectors”

0:00:04 · Daniel Faggella · Podcast history

The AI in Banking podcast runs four in-depth interviews each month with bank leaders, venture capitalists and vendor AI experts.

“go into four in-depth interviews every single month with leaders within banks, with venture capitalists, with AI experts at various vendor companies”

0:01:11 · Daniel Faggella · Podcast format

Ian Wilson is a consultant in AI and banking with a startup background who was previously head of AI at HSBC.

“Ian is now a consultant in the AI and banking space. He has a background in the startup world, but he was previously the head of AI at HSBC.”

0:01:44 · Daniel Faggella · Guest background

Emerj's AI and banking vendor landscape and capability map report examined all AI banking vendors and dozens of internal AI use cases at leading global banks.

“essentially where we looked at all AI banking vendors and dozens and dozens of internal AI use cases at global leading banks”

0:02:14 · Daniel Faggella · Research report

Ian Wilson was one of the formal research advisors for Emerj's AI and banking vendor landscape report.

“Ian made a big contribution by being one of our formal research advisors for that report.”

0:02:48 · Daniel Faggella · Research report

The podcast's next monthly series, in July, was to focus entirely on fraud, risk and compliance.

“Next month, for example, in this coming month here of July, is entirely focused on fraud, risk, and compliance.”

0:02:48 · Daniel Faggella · Podcast schedule

Ian Wilson says banks went through a similar pattern with mobile banking, funding many projects that went in different directions.

“I think the same process happened with digital and banking with mobile in that you would have money thrown at all sorts of different projects who would shoot off in all sorts of different directions.”

0:07:24 · Ian Wilson · Technology adoption

Ian Wilson describes a chief AI officer as a group-wide role that defines what each business line should do with AI and how to carry it out.

“this group-wide role whose function is to strategically define what each business line should be looking at in terms of its AI strategy and how to execute those.”

0:08:41 · Ian Wilson · Chief AI officer

Ian Wilson says operations, retail, commercial and private banking each have different strategies, so each needs a different AI strategy.

“And operations has a completely different strategy to retail banking, which has a different strategy to commercial banking, which has a different strategy to private banking”

0:09:09 · Ian Wilson · Business strategy

Ian Wilson says AI is a very large IT deployment, not an application that can simply be installed and run.

“You don't just download it onto your desktop and start running it. It's a very large IT deployment.”

0:09:42 · Ian Wilson · AI deployment

Ian Wilson gives a hypothetical commercial banking strategy of recommending better products to existing customers and finding new customers through them.

“our strategy is to recommend better products to our existing customers and grow our customer base by finding customers that are already doing business with our existing customers.”

0:11:00 · Ian Wilson · Commercial banking

Ian Wilson says commercial bankers currently use Excel sheets drawn from their data stores to decide which companies to target.

“commercial bankers will take Excel sheets from their data lakes or whatever and try and prioritize who would be a good company to go after.”

0:12:13 · Ian Wilson · Commercial banking

Ian Wilson likens an AI tool that surfaces target companies from the bank's data to Facebook for banking.

“you could imagine an AI capability being built to do exactly that because it's essentially Facebook for banking”

0:13:12 · Ian Wilson · AI use cases

Ian Wilson expects AI products, vendors and teams to mature dramatically over the following five years.

“a key part of AI is going to be its roadmap over the next five years. It's going to mature dramatically.”

0:14:32 · Ian Wilson · AI maturity

Ian Wilson says much banking work consists of reading information from paper and typing it into databases, which makes it a target for automation.

“there's so much paper in banking. And so much of the work is just read some information off of a piece of paper and type it into a database.”

0:15:31 · Ian Wilson · Document processing

Ian Wilson says reading a million documents in an hour rather than a month would enable new uses a business could not previously imagine.

“If we can read them in an hour rather than in a month, then we can start doing all these new things we couldn't imagine doing before.”

0:16:07 · Ian Wilson · Document processing

Ian Wilson identifies business comfort that an AI tool will work as one of the biggest hurdles to deploying AI in banks.

“one of the biggest hurdles I've seen to deploying any AI solution has been the comfort factor.”

0:19:11 · Ian Wilson · AI adoption

Ian Wilson says doubts about an AI tool disappear once a vendor can show that competitor banks already use it.

“well your competitor banks x y and z are all using it, then the question disappears.”

0:19:37 · Ian Wilson · AI adoption

Emerj's research sees AI adoption in banking unfolding in three phases over roughly six to eight years.

“We sort of see this as three phases moving forward over the course of about six to eight years”

0:21:16 · Daniel Faggella · Adoption phases

Dan Faggella says the current phase of AI in banking is risk reduction, with efficiency following as phase two.

“The current phase of AI and banking, interestingly enough, is not efficiency. We see that as phase two. The current phase is more about risk reduction.”

0:21:16 · Daniel Faggella · Adoption phases

Emerj reports that well over half of the vendor funding raised across the banking AI landscape it analyzed is in risk, covering compliance, fraud and cybersecurity.

“Well over 50 % of the vendor dollars raised across the banking landscape for all the vendor companies we analyze is in the domain of risk.”

0:22:18 · Daniel Faggella · Vendor funding

Nishant Chandra, the next episode's guest, is head of data products at Visa in Bangalore.

“He's the head of data products at Visa in Bangalore.”

0:22:45 · Daniel Faggella · Next guest

Chapters

  1. 0:00:04Series launch and what the show covers

    The host introduces the new AI in Banking podcast, its monthly format and its audience of senior business leaders.

    “AI in industry has been around now for four plus years”
  2. 0:07:49From experimenting to structure

    Wilson says banks spent freely on AI with little structure and now face questions about return on investment.

    “furtive money going left, right and center, but no structure to it”
  3. 0:09:09Each business line needs its own AI strategy

    He argues that AI is a large IT deployment that must be matched to each business area, not a single bank-wide plan.

    “your business lines need an AI strategy”
  4. 0:11:42Two phases: improve first, reimagine later

    Wilson explains that early AI should improve current work, and business teams will reimagine their operations once they see what it can do.

    “The first is do what we do today, but better.”
  5. 0:15:31Boring automation delivers the first returns

    Paper-heavy processes are the low-risk place to start, and the learning gained opens up new uses.

    “a very boring but very massive return on investment for a bank”
  6. 0:19:11Learning and comfort drive adoption

    The business must understand the capabilities and trust the tools, and proof from competitors would settle the doubts.

    “One is the learning, but second is the comfort factor.”
  7. 0:21:16Research view: risk comes first

    The host says the current phase of banking AI is risk reduction, not efficiency, and where vendor funding is concentrated.

    “The current phase of AI and banking, interestingly enough, is not efficiency.”

Transcript

0:00:04
Daniel Faggella

This is Dan Faggella and welcome to the first episode of the Emerge AI in Banking podcast. I'm excited to kick off this series and I'm glad that you're here with us on this first episode. Many of you may already know us for AI in industry. AI in industry has been around now for four plus years where we interview AI experts in various sectors to essentially get cutting-edge use cases and current trends on where AI is disrupting pharma, insurance, banking, defense, and find those transferable lessons for leaders who are looking to apply AI in their own business.

0:00:36

Not technical folks who write the code, people who command the dollars, command the initiatives, command the strategies. That's our audience. And now with AI and banking, we're focusing specifically on the banking sector. We have tens of thousands of banking folks reading our material at Emerge.com every single month, and we wanted to dive in with a specific podcast to serve their needs, to explore where banking is headed. So this AI and Banking podcast, what I'm going to do in this intro, which will be a bit longer than the ones in the future episodes, I promise the other ones will be much shorter, is go over what this podcast is going to be like, what you can expect as a listener

0:01:11

here, as well as what you're going to get out of this particular episode and the coming episodes in this short series, this five-episode series. So the AI and Banking podcast is broken out like such. We go into four in-depth interviews every single month with leaders within banks, with venture capitalists, with AI experts at various vendor companies who are servicing the banking sector to look at a particular theme. And this first series of five episodes, which we're releasing all at once, we're releasing five episodes to kick off this series, are all about the future of AI in banking.

0:01:44

So where is AI taking us forward looking? That's the theme, the thread across these five initial episodes. And this first episode, we're joined by Ian Wilson. Ian is now a consultant in the AI and banking space. He has a background in the startup world, but he was previously the head of AI at HSBC. In terms of financial institutions, it quite literally doesn't get bigger than HSBC, a company with massive resources and also massive challenges to overcome when it comes to applying artificial intelligence.

0:02:14

So Ian brings a diverse background in terms of independent work, the startup environment, and the big banking environment and brings all of that to bear in this particular episode here. So he's going to be the one kicking off the series. Ian is also one of our formal research advisors. So many of you have seen the launch of our artificial intelligence and banking vendor landscape and capability map report. This is essentially where we looked at all AI banking vendors and dozens and dozens of internal AI use cases at global leading banks to find the trends of where AI is making the biggest difference now.

0:02:48

Ian made a big contribution by being one of our formal research advisors for that report. And that's why I wanted to be able to really lead off with some of his insights here, which I thought were very prescient. And I should note that in every series that we do here, when we do these four interview series, which every single month we have a new theme. Next month, for example, in this coming month here of July, is entirely focused on fraud, risk, and compliance.

0:03:12

So the episodes that will be dripped out every single Monday for the rest of July will be about fraud, risk, and compliance as an end cap to all of our episode series. So this current series on the future of banking, the next series on fraud compliance, and all future series that we're going to be recording. At the end of those four interviews that we do every month, we have an analyst deep dive session. And what we do here is we essentially talk to members of the Emerge Artificial Intelligence Research team about the critical takeaways from those interviews and how that meshes with the data that we're collecting for our own research.

0:03:45

So this is kind of an opportunity to look behind the curtain, get a sense of some of the facts, the figures, the ideas, the core numbers, the percentages, the dollar figures of where AI is headed in banking and how that correlates to what our experts have said. So you'll get the qualitative insights, so to speak. And then in the analyst deep dive, you'll also get to peek at the quantitative insights. And that's going to happen at the end of every monthly series.

0:04:09

And this initial five series of the future of AI and banking is no different. In fact, you'll be able to see that analyst deep dive as our fifth episode. And you'll get one at the end of July as well. So that's what you can expect moving forward, diving in with Ian. What I should mention here are some of the major points you'll learn from this particular episode. So Ian talks first about how each line of business and banking requires its own AI strategy and why it's a little bit ill-advised to ask what is our AI strategy for a whole bank.

0:04:38

Ian goes into some detail that I think is very useful for leaders in banks as to why individual business units, lines of business, require their own AI strategy. He also talks about the phases of AI adoption and banking and how the early phases are really all about improving what we do already, being able to build competence and improve systems rather than necessarily having to rethink banking wholeheartedly when it comes to artificial intelligence. It's a bit too much of a reach for banks.

0:05:05

Ian talks about a more gradual core capabilities approach, and we'll dive into that in our episode with Ian. And in addition to that, he talks about how important it is for business leaders to understand the use cases of AI in banking and how banking leaders should be the ones that ultimately lead initiatives. That initiatives should not be driven by IT. They should not be driven purely by data scientists. They should really be driven by business leaders who see opportunity.

0:05:30

And Ian talks about how to enable that within banks. Again, this is a fellow who is the head of AI at HSBC. So some pretty grand sort of experience in terms of being able to look at how an enterprise views this. And I think it'll be useful for essentially everyone tuned in here as well. So without formal ado, we'll fly right in. If anybody wants to see the full report that Ian contributed to, the AI and banking vendor landscape and capability map report, we actually have a executive summary that's downloadable where you can get many of the best charts and graphs from that entire research.

0:06:00

And that's just at emerj.com slash b1. That's b as in banking and then the number one. So emerj.com slash b1, emerge slash b1. You can just download a sample. I think it's 10 or 11 pages of some of our key insights, some of our key tables and charts to get a sense of what was in the full report and to learn some cutting edge insight that I hope will add to this episode. So I'm Dan Faggella.

0:06:25

I'm joined here by Ian Wilson. And without further ado, let's fly into this episode of AI and banking. So Ian, the first interview, I'd like to kick off with sort of your thoughts on the biggest high-level lessons and understandings that the boardroom and the C-suite in banks should understand now when it comes to AI. What to you really jumps out as those big bullet points?

0:06:56
Ian Wilson

So you probably talk, I would imagine, around very specific applications, but I would probably start by taking a couple of steps back. I think one of the key trends I've seen over the last couple of years has been learning. There's a lot of learning going on right at the C-suite and understanding about what AI is. And I think we're seeing a transition from the let's just jump in and do something cool phase to actually let's do this correctly phase.

0:07:24

And what I mean by that is, and I think you'll see this as a repeating pattern you see with any emerging technology. I think the same process happened with digital and banking with mobile in that you would have money thrown at all sorts of different projects who would shoot off in all sorts of different directions. People start building groups within the AI sphere. Lots of different groups within any particular bank would have had all sorts of different AI groups.

0:07:49

And what you're seeing is perhaps not the ROI that you were expecting, but you are also seeing a bit of chaos in that you've got all this furtive money going left, right and center, but no structure to it. And I think the key question you'll be hearing at the C-suite within a bank particularly is what's next? What comes next? As well as where is the ROI? And often there's no answer to that question. There is no what's next because there was nothing strategic put in place.

0:08:18

It was just throw some mud at the wall and see what sticks, which is fine to learn. But what I think you will see transitioning now within the C-suite is the recognition that, OK, we need to put some structure around this. We need to do this strategically. And where I see that shaping up, and I've seen a couple of indications of this, is I wrap that in what I call the role of a chief AI officer.

0:08:41

Now, it doesn't have to be a chief AI officer, but I think that encompasses it nicely, which is this group-wide role whose function is to strategically define what each business line should be looking at in terms of its AI strategy and how to execute those. And where I say each business line, it's another key point and another key learning is I hear often and I read often this term about what's the AI strategy. You need an AI strategy.

0:09:09

Yeah, yeah. Well, you don't. your business lines need an AI strategy. And operations has a completely different strategy to retail banking, which has a different strategy to commercial banking, which has a different strategy to private banking, which has a completely different strategy to investment banking. And I think that is what I think will be the next phase, that growth and maturing into that level of understanding that actually AI is not just an application. It's not Microsoft Word.

0:09:42

You don't just download it onto your desktop and start running it. It's a very large IT deployment. It requires a lot of thinking, and it requires some strategic thought about how we align it with each different business area. I think for me that really is what I'm expecting to see, and I'm starting to see across peers in different banks.

0:10:06
Daniel Faggella

Yeah, so let me just kind of put this in a nutshell, Ian. So you just dropped a lot there, and I think these are important high-level ideas. First and foremost, certainly, yeah, looking at individual applications I think would be maybe the most wholeheartedly horrible way to focus on AI first within banks. a lot of the consensus in the report is getting the foundation right. And part of the foundation that you're addressing here and putting a lot of emphasis on is kind of a higher level strategic perspective.

0:10:32

So not treating it like a bunch of toys that we'll let people run with, and maybe not treating it like it's just IT, but like it's a new set of competencies that need to, you know, have some kind of central hub for how they'll be rolled out, where they'll be determined and where they'll be prioritized and that that needs to splay differently across different parts of an organization as large as a bank. Is that an okay summary here or do you want to kind of put a

0:11:00
Ian Wilson

caveat on that? No, I think that's exactly what it is. For me, it all starts with a strategy. So if you take, let's say you take the strategy of commercial banking, for example, you look at, well, the first question you would ask is what's our commercial banking strategy? And if you're building out an AI strategy, it's not an independent item. It should be something that feeds into and supports your business strategy. So if you imagine commercial banking, for example, they might say, well, our strategy is to recommend better products to our existing customers and grow our customer base by finding customers that are already doing business with our existing customers.

0:11:42

then you might think, okay, what capabilities can we build? What capabilities can AI offer to enhance that strategy that were never available before? And I think you'll find very much there's two phases of deployment for emergent technology. The first is do what we do today, but better. The second phase is reimagine our business. And for me, the people that will reimagine your business are the business themselves. It's not the AI people. I'm not going to go and reimagine commercial banking.

0:12:13

But once the commercial banking group have used AI capabilities and started to see what they can do, what they can offer, they themselves will start to reimagine new ways to use that technology. But I think that will be phase two. Phase one is just do what we do now. And to my example I just gave there, I know, for example, commercial bankers will take Excel sheets from their data lakes or whatever and try and prioritize who would be a good company to go after.

0:12:43

Maybe they find a company that's already trading with 50 of our current customers. Okay, we might think that's a good one to go after. But that's done today manually using Excel. What they would love to see, and I've had conversations with this, they would love to come in in the morning, open up an application. And those top companies have been raised and pulled out of the entire corpus of data that's in the bank to say, focus on these companies because that's all the business needs.

0:13:12

But to get that information is currently not possible in any practical sense. But you could imagine an AI capability being built to do exactly that because it's essentially Facebook for banking. That's the kind of thing that that application is directly driven by the requirement to support that business line strategy. And then that's where you drive your AI. And how do you build that? That really is

0:13:38
Daniel Faggella

then a next level that you take a look at. Yeah. And I think the idea that right now, AI is being seen not as a industry re-imaginer, but as kind of an enhancement or augmentation to existing processes, I think very much to me feels like the paradigm within a great deal of finance, nevermind banking. The money is being spent on these kind of more defensive applications. We're not trying to find some new user experience here or some great new way to reach new customers.

0:14:09

It's kind of like, can we improve a little bit of efficiency? Can we add an extra level of capability to RPA or something like that? What do you think it is that gets an industry, a sector to kind of turn the corner from being in a space where mentally we're really only thinking about incremental and efficiency improvements to finding new opportunities in a

0:14:32
Ian Wilson

bigger picture. What makes that happen? I think, again, it's knowledge and lower risk. So the key point about AI for me is its risk. It's highly uncertain. So there's a lot of risk with any deployment. And as we go on through time deploying these solutions, the comfort within any organization, within any enterprise starts growing because you can see, you'll start to see things delivering ROI. And also, as we move down the road, a key part of AI is going to be its roadmap over the next five years.

0:15:05

It's going to mature dramatically. The vendors are going to mature dramatically. Our product's going to mature dramatically. Teams who deploy this are going to mature dramatically because they're going to start learning how to do it. They're going to start learning how to do it repeatably. They're going to have cookie-cutter models for their operating models. But we're not there yet. So where we are today is lots of uncertainty. And with any kind of emerging technology like this, where is the first place you deploy that?

0:15:31

Well, you look for your low-hanging fruit. And the low-hanging fruit is the stuff we're already doing now. So an example within banking is there's so much paper in banking. And so much of the work is just read some information off of a piece of paper and type it into a database. It's like plumbing. There's nothing exciting there. But if you could replace that reading off of paper with using some vision processing, machine learning, natural language processing to do that automatically, that would be actually a very boring but very massive return on investment for a bank.

0:16:07

And that's the kind of place where I think you'll see it moving in. And then, again, the business starts to learn and say, oh, well, you know what? If you can read from that document, doesn't that also mean you can read from this document? And then the next step for a business is understanding, oh, wait a second. So, you know, we've got a million documents coming in. If we can read them in an hour rather than in a month, then we can start doing all these new things we couldn't imagine doing before.

0:16:33

And I think that comes from learning, understanding the capabilities of the technology. So for me, technology, natural language processing, blah, blah, blah, I never talk about that to the business. I always talk about what is the capability it offers you because that's really all you care about.

0:16:50
Daniel Faggella

We call the report the capability space for that reason. Yeah, saying here's where anomaly detection, right? Looking at a level of AI approaches, who cares? But give me the new verbs that I can do in banking. right? I mean, that's the whole lens of the research for that reason. And so it sounds as though you're of the belief that that knowledge of those capabilities and a little bit of experience with some of the low hanging fruit combined with the vendors maturing and making these technologies more accessible and more capable, those things will kind of create the blossoming of brand new ideas, brand new ways to take and apply these technologies that maybe those will be kind of

0:17:28
Ian Wilson

enabling factors? I think so, because I think that there's already, strangely, I found so much enthusiasm for AI in banking, which is completely based on really not understanding at all what it does. And it's a shame for that to dissipate. And I think as long as we can start delivering some ROI, then we can keep that enthusiasm up, but then start transforming it into knowledge and combine that enthusiasm with real knowledge and have those. What you can get then is the business driving the need.

0:18:01

So rather than having an AI group or a technology group trying to say, we've got this cool stuff, can we use it? You've got the business saying, we want to do this new thing. Give us the capability to do it. And I think that's an important driver, which is why I always stress that the learning and having the business drive this is the key factor because often they have the budget too, but they're the ones who will say, you know, we want this as opposed to an IT function or a technology function trying to push

0:18:30
Daniel Faggella

that, which I think is the wrong approach. Yeah. And I guess, you know, to your point, a business wing would need to have enough of a, let's say, non-naive, non-ignorant perspective on AI's capabilities, what's realistic, what's not realistic for them to come up with those ideas. So it's this kind of lengthy education process. It sounds like you're kind of agreeing with the quote that I realize I probably overuse, but I still like it, which is that most of the money in AI is in events where everybody goes and gets these ideas, right?

0:19:03

It's not from people in the business arm saying, oh, my goodness, I need to solve this problem or do this. It's people figuring out what the heck is going on.

0:19:11
Ian Wilson

Well, it's a problem with the business right now. Again, it's twofold. I think the second part shouldn't be underestimated and really should be overstressed. One is the learning, but second is the comfort factor. Many times, one of the biggest hurdles I've seen to deploying any AI solution has been the comfort factor. And that's just getting the business comfortable that this thing is going to work. And that's a key challenge for most vendors because they can't prove it.

0:19:37

They can't show where it's already in production. because it's difficult particularly with machine learning to really give you very simple clear answers and that's a big stumbling block right now but we'll become i think completely moot once we just blow past it and actually get things in production because really that the proof is yes it's you know if somebody says oh yeah we like this idea of this this tool you've got but we're not sure if you can say well your competitor banks x y and z are all using it, then the question disappears.

0:20:11

And I think that will be another key factor in the maturing is once it gets out there and people start using it and it's producing ROI and it's known to produce ROI, I think a lot of these questions go away.

0:20:25
Daniel Faggella

And we are not there yet, Ian. But hopefully in terms of the, hopefully some of the audience has a better perspective on the high level by hearing that high level from you. Obviously, those of you who are tuned in, we have two more interviews with Ian's deeper takes on the full-blown report. But Ian, that's all that we had for this interview. I sincerely appreciate you sharing the big picture with us on the show.

0:20:46
Ian Wilson

You're quite welcome.

0:20:51
Daniel Faggella

So that's all for this episode of AI and Banking. At the end of each episode, what we like to do is summarize sort of one of the key points of what our expert interviewees had said and how that ties to what we've learned in our research. So after every episode, you'll get kind of a research nugget in addition to a preview of the upcoming episode. And so with Ian, one of the important threads that he was touching on are the phases of AI adoption in banking.

0:21:16

And our vendor landscape report of AI and banking goes into phases in great depth. We sort of see this as three phases moving forward over the course of about six to eight years, at least by our present estimates and by dozens and dozens of interviews in this space. The current phase of AI and banking, interestingly enough, is not efficiency. We see that as phase two. The current phase is more about risk reduction. So many of the low-hanging fruit AI use cases in banking have to do with reducing risk.

0:21:44

Fraud, compliance, cybersecurity, essentially detecting anomalies. That's not nearly as hard as having AI write new messages to a wealth management client to get them to invest more money or something like this. There's many more complex speculative AI use cases that are in the works, but places that already have a lot of automation efforts and already have a natural fit for AI in many respects is risk, not to mention bankers are focused on risk. So to kind of marry with Ian's insights about the phases of adoption, one of the things that folks that have seen our full report will know is that we break that down into three phases.

0:22:18

And I think it's important for banking leaders and for vendor leaders to understand that really where a lot of the attention and also where a lot of the money is focused when it comes to AI and banking is on risk. Well over 50 % of the vendor dollars raised across the banking landscape for all the vendor companies we analyze is in the domain of risk. So compliance, fraud, cybersecurity. In our next episode of AI in Banking, we'll be interviewing Nishant Chandra.

0:22:45

He's the head of data products at Visa in Bangalore. I met with Nishant myself when I was in Bangalore last. And in our coming episode, Nishant talks about an analogy for AI adoption in banks that involves Italian food. That's my only clue for you right now. Nishant is a man of many analogies. It's hard not to like the guy. Incredibly pithy insights on AI adoption in banking. This is someone who served in high-ranking roles in Visa and AIG and other large finance businesses and has a good sense of how AI is making its way into banking, integrating itself into banking, a very clear view of sort of how that will shape forward into the future.

0:23:20

And again, that is the theme of this five-interview series here on AI and banking. So Nishant will be up next. I hope you join us then as we dive into all of his insights as well. Thanks for being here on our first-ever episode, and I will catch you in the next episode. Okay.