Courtesy of Dell's Bill Wong below is a transcript of his speaking session on 'AI-Powered Digital Transformation in Financial Services' to Build a Thriving Enterprise that took place at the Business Transformation & Operational Excellence Summit in Financial Services Live.

Session Information:

AI-Powered Digital Transformation in Financial Services

AI is fundamentally changing how businesses operate, redefining the way people work and transforming all industries worldwide.

In pursuing digital transformation initiatives, organizations are seeking how best to leverage data as an asset, implement a data-centric culture, and capitalize on machine learning algorithms to improve their customer experience and improve their competitiveness.

This session will focus on best practices for digital transformation initiatives and how AI and data analytics can be leveraged to enhance Business outcomes. The latest trends and direction of AI technology, key challenges, and recommendations will be reviewed.

Topics include:

• AI in Financial Services
• AI Challenges
• Leveraging Data as an Asset
• AI Best Practices

Session Transcript:

Bill Wong is an AI and AI and data analytics practice leader at Dell technologies in Canada, as 20 plus years of experience in AI and analytics, working with only some companies like IBM, Microsoft, Oracle, and Dell Technologies. In other words, he's in the vortex and the way all this stuff is moving around.

He's published numerous books, which is fantastic to anybody who knows how hard it is to get a book out the door. And white papers, which are even more useful, and shorter and easier to read. And one scientific paper on a fascinating subject of: How do you use drug, discover how to use AI, and machine learning to do drug Discovery, which you, you may know, is an incredibly complex and time consuming task.

How can that be done better with Machine Learning, and a quarterly AI newsletter distributed worldwide?

So, without further ado, I'd like to turn this over to Bill, who tells me, this presentation contains a lot of brand-new material ideas, and we all get to be the first audience to hear it.

Thank you very much for joining us, Phil. I'm going to disappear into the background, but I will be here throughout the session, and I'll be looking for the questions that come in for the Q&A at the end of the session.

See you soon.

Thanks very much, Chris.

Hi folks. Morning, good afternoon, good evening, wherever you are.

And yes, welcome to this session. I'm going to talk about AI in financial services.

And you get bonus marks if you can recognize the icon on the bottom right hand corner.

That's the city I'm presenting out of, and that's Toronto.

And I was with it.

the GM of one of the nvidia divisions looking after financial services.

And he said to me yesterday, said, well, Canada, it's great to talk to you.

You guys certainly hit above your weight, And then, and that's how Canada is kind of perceived in AI.

You've been here before.

You'll know and Nvidia, and other companies refer to, as tried to, as kind of the the first place, or the big bang of the latest generation of AI, the new AI here.

It was the first time, and there's a contest called imagenet, where we use GPUs, deep learning algorithms. We have lots of data.

And that spawn papers, it spawned, or the Turing award awarded to a number of the researchers who are involved there.

And since then, that was back in 20 11. And AI has been continuing to grow like gangbusters and probably no other industry.

No, other industry, financial services, it leads in the adoption of AI, and we'll take a look at some of the reasons why there.

So, it's cool, So, talk, high level about the environment that we're all working in.

I'm going to focus on AI and automation tools.

And finally, I'll talk about the challenges and end on a positive note that, yes, there are lots of things that we're doing to try to address these challenges, deploying AI.

So, the environment that we're in, the opportunity, as the practice leader for AI, to, to deal with a lot of different organizations in financial services.

And especially the banks, the banks are quite concerned about the competition.

The emergence of the fintech, we'll talk about them, but not just the fintech's.

You have the car, digital commerce giants like e-bay, Amazon, and their own payment schemes. If you do e-bay, you know that you're using PayPal.

And a lot of them talk about what happens.

Well, what would happen if somebody like Google entered financial services?

I know a bunch of insurance companies, they ask themselves that.

But from my perspective, I don't think that'll happen again that that's my opinion, certainly they're gonna push the envelope and try to get what business they can get like like payments, for instance or credit cards.

But once you join, you know into the fray of Financial Services, what happens is on the right a massive amount of regulation is required.

And that doesn't really jive well with any of the hyper scholars They, they probably have been, I guess, under the radar when it comes to regulation with all the governments around the world.

And that's not going to change, just because they enter financial services.

So, but I do think it's useful to use them as a benchmark.

And why is it useful to use as a benchmark, is the bottom left-hand corner, this change in market.

Everybody wants a piece of the next generation coming out.

Be it Gen Z, the Millennials, whatever you call them, these are digital savvy people.

And the interfaces they like, no, they, they gravitate to the Digital Commerce Charts and hyperscale, not traditional financial services companies.

On the far right, lots of, uh, huge, macroeconomic trends that are occurring.

Of all of these things that are placed here, probably the only thing that you have, some level of, control, is what technologies you can pick, too, to thrive in this environment.

And so, I'll talk about AI, automation, and Cloud briefly, but hopefully give you a good overview of these technologies, and what you can do.

Now, the millennials, very attractive market, my favorite kind of observation about them is this last one here. That 71% of millennials would rather go to the dentist and to listen to what banks are saying.

So, don't take as a slight millennials don't like to listen to anybody, but especially kind of your, your, your father's or mother's financial institutions. They want something that is at the palm of their hand.

And what's coming up?

I don't know if it's used a lot, but people are starting to talk about the Gen C, some people called this, the connected consumer.

Some people are calling it little, kinda, the code, the generation.

That will have a mark on these people. That's kinda the next generation that's coming.

But, again, the ones that are gravitating toward are Google and Amazon.

And in the midst of this, as we've tried it, select technologies, that will give us a competitive advantage in the marketplace.

This metric here, that there's a huge shortage of IT skills.

That's nothing new.

And, I submit that. It'll always be like that, until perhaps, maybe use AI that you won't need any developers. But, we're, we're far away from that right now.

Um, and, in some surveys, some HR professionals are seeing that, up to 70% of the developer community.

It's either open or actively looking, or a change.

And the reason for that, I don't know if you know any developers, but I've worked with developers my whole career.

And I find that, know, developers don't stage developers, you don't see too many. People are like 20 years.

Oh, developing the, it's, it's, it's an industry. It's a, it's a area of technology that.

No, they're not waiting for the next programming language to learn. It's, it's a different there.

So, typically, developers, what happens, What I've seen as they may start their own firm, they might do consulting, that would become solution architects.

But it's rare that they say, hey, after 10 years, I still like to code.

So, they, they, they change and I think we'll always be challenged.

So, what do we do in this environment to move these folks to really adopt these new technologies going forward?

So, enter the business technologist, a number of companies, including Gartner have talked about this group, and I'll bet you people who are watching this right now. Some of you are what we would consider a business technologists. Now, this isn't a title.

It's people who are not reporting to the IT department, but are technology savvy.

In this chart here, you'll see that, you know, only 10% are really officially IT type staff, with only 4%.

Officially in IT departments, 6%, even more.

We've seen a lot of business units, hire their own IT staff, because they can't really depend on central IT to to get what they really want.

Then the rest, 49%, They're happy just to receive, you know, what the IT department sends them.

Now, this 41%, she just mentioned, this is kind of an average in some industries, such as government, this will be a lower number, about maybe 25%.

In my industry with technology, there's probably around 50%, and I would submit that in financial services as well, it's, it's higher than 41%.

So, these business technologists, types of business technologies that we're seeing emerge, are what we call citizen developers, citizen data scientists, or citizen integrators.

Again, these are folks who report to the line of business and not officially IT, but they know how to use the tools.

If you take a look at the bottom part of the graphic here, they re to cleanly use Automation, integration, application development.

And data science tools.

Without the aid of IT, except for perhaps me to install it.

Now, the reason why they're using these tools, typically, is to find insights, Customer insights, could be product insights, but they want to better understand how something's working, why it's GVHD.

Or another popular area is they want to optimize the business process, or process, or south of the border, automate tasks, or to integrate the customer, experience the customer journey.

So, it looks like you have a 360 view of the customer. And your customers are dealing just with one firm.

These are the most popular types of projects, and those projects drive these kind of tools.

Now, the tools that I've listed, there was a recent survey, and I've actually list the tools and kind of the order of its popularity.

So, in Data Science, AI tools, the most popular tools that people are using, our data natural, or, sorry, or database tools, things like SQL Server, instance, BI, tools like Power BI or Tableau, as you get down to the lower level tools.

There, they become a little more sophisticated, a little more complex, although there are tools there that I believe salmon users, business people could use it, or, it's still kind of early in business. Process automation.

Lots of tools out there where you don't need to be a developer, No code required there.

Same with the automation tools.

These are really, more and more, they are really geared toward business, professional, what understands, how the business functions, rather than IT people.

And this slide here.

What is China's ship?

Convey is these skill sets, their relative strengths or weaknesses of each of the skills that are listed on the far left?

The launch of the data ability to manage data, business acumen, coding skills, et cetera. Now, what I put it in the red boxes are the citizen, integrators, the citizen data scientists and citizen developer.

So I've been called citizen data scientists because I tell people quite openly I hate to code.

I mean, there are sometimes, you know, I might do it just for fun, but I don't like classical coding.

Well, I take data, put it into some kind of program, do some analysis and analytics, I'm all in there.

And I would say that this probably is somewhat accurate.

I would say my data skills are more like a data engineer, So probably it higher than the average citizen data scientists are.

If you look at the column, a couple over, the data scientists, you'll see for coding skills.

Deep into Python.

Deep into machine learning frameworks there. And that's something that's interesting, though.

I don't feel, let's say that that that's the funnest part of being in this data science role.

So I'm happy to give that to the teacher scientist.

The forecast is that, you know, by 2023, we're going to have four times as many of these folks.

Then then your classical developers here.

So let's talk about, little about the tools here.

So, when it comes to AI automation tools, the, the popular category that we have, and I've been evolving this chart, um, through the last couple of years.

Business process modeling, well understood technology has been there for another years.

Focusing on, on the business process model, RPA.

Robotic process automation, very good, in terms of getting an ROI, so low hanging fruit, and we'll talk about how that's evolving.

Process mining is relatively new.

It's exciting, new technology and what this does is, it compares your existing workflows process to be optimal and then identifies, shows you deviations of where you can improve.

Then, finally, AI, Charles spend time on the challenges.

What it's trying to do here is really simulate human decision making.

Yeah.

That can be a challenge in some use cases.

So, going AI and automation tools adoption.

The first one, a big kind of vote for process mind that that's an up and coming technology.

But the other two kind of forecasts I'm seeing, as well is that AI, whether you're gonna know it or not, it's going to be integrated in all the tools you use.

Like, for instance, HR, HR, people are asking questions like, who's likely to lead within the next 12 months, You know?

And if that person has a title of developer, that's probably one of those job types. But outside of job titles that are there are certain activities, engagements.

tells things that employees communicate.

The programs out there are our monitoring, right.

Gives people a better idea, Hey, perhaps there's something you could do here, let's talk to the person, maybe enrich the job, Maybe compensation isn't what it should be, so, whether you know it or not, or you code it, AI is going to be embedded a lot of their future tools.

Now, we get to financial services, and folks I've talked to, I haven't really found too many areas of financial services that AI cannot impact.

Whether you're in the front office for back office, in middle office, There are opportunities there.

Now, where I see the large focus on one of the banks, I would say three quarters of their AI projects really are on the far left here is they want to tailor that customer experience make their products easier to understand.

Um, of course, they still want to be leaner.

The cost is always a great way to costs justify, But that's not the biggest driver that I'm seeing right there.

And when you talk to fintechs, you know, they're innovative group.

And they're looking for new types of products.

So, a couple of examples here, what AI enables you to do.

Personalization. But really, you know, the, the Holy Grail is to have really so great personalization.

And it's hyper personalization.

Now, think a lot of folks are there yet.

There are quite happy just to have a chatbot that that is friendly.

one instance, Kapil one is, it's cited as one of the first banks in the US.

To come out with their own kind of tailor chatbot, and what they decided to do is it's kind of interesting.

They have a personality that they call, if the name I think it's no.

No, and, it's, uh, and backwards it spells one, I don't know if that means anything, but it's vendor neutral, chart, gender, neutral, and you can tell it to automate your payments. Get alerts for possible fraud activities.

But, you could also ask it questions like, what is the meaning of life?

So, they're trying to make this a compelling user, experience that it's more than just automated bought to do financial transactions.

So that's interesting strategy there.

What we're seeing with automation is augmenting traditional automation with RPA.

So JP Morgan has an application called coin.

Queen stands for Collective the Contract contract Intelligence, SEO, contract intelligence.

And they have activities every year to make sure that they comply well, and they have a task that these measure, that takes 360,000 hours involves lots of professionals, lawyers, et cetera.

And they claim with AI, there's an RPA and then with natural language processing, image processing to lead the contracts.

They've reduced that time, two centers.

So you can read about that though, however, yeah, a comment on that or a description of that in the appendix on that.

Then finally, risk compliance, especially for things like fraud detection, very popular financial services.

Right?

Comes to mind mastercard's.

So, couple of years ago, I was speaking at sea, Ottawa, the capital of Canada.

And the person came up to me and said that.

Our key, he drove in from Montreal, and he used his credit card just before leaving.

Then he stopped her gas, and then the credit card was denied.

And she called the credit card company, said, Hey, you got this false positive.

Like, well, why, Why did you do that? Like the coach she tell like Jess used it.

And the person on the phone said, Well, you did user, Yeah, we'll see that transaction, but where are you all right, now? There's, There's no way. You should be at the place that is given time.

And he just laughed, and if you've ever been to Montreal, you know, these folks who drive, like, there's no tomorrow speeding down the highway.

So he was there in, at a time, they really should have been.

So you laugh that I said, OK, that's that's a great help.

I'm OK with that. False positive.

So let's go.

Now enter the fin techs.

All the traditional banks, I talked to, you know, are concerned about this group.

Because they don't have the legacy, there are recent grads, all the latest ideas outside of that, If you take a look at what they're developing, on this left hand of this graphic here, for the larger syntax, they use AI to develop new products and services.

As opposed to the incumbent, the traditional type of banks.

Only 15% of their deliverables are considered, like, AI type enhanced products.

So, one example is that, is, a lot of these companies will sell to the algorithms to do loan processing.

So then, you know, based on what they have, they have the latest, greatest state-of-the-art leverage, new types of data, avoiding biases, and these kinds of things.

And so successful, I guess, that they're willing to sell that.

Now, most traditional banks that I know would not do that.

There no intelligence, they built up, they want to keep that.

So it's an interesting way of how these couples race.

Just going to say, there are lots of challenges, even though AI has the possibility of transferable and really changing the way you do business.

It is fraught with lots of challenges and gardeners, you know, they're estimate 85% of those projects.

So, why, why is that?

Probably from a technology perspective, the number one reason is has to do with the data, lot of times, Nevermind the data quality.

Every lot of folks can get at that kind of customer data that they need, but even if they can't get some approximation data, trying to make it clean, usable, audits, et cetera, near real-time, these are challenges when you build AI applications.

And, the best practice with any kind of an AI endeavor is first to start off big picture, have a strategy in place, know why you're in the scheme, know, understand why you want to use AI in this area. What is kind of the end goal?

that allows you to pick the right use cases support to pick your technology roadmap, but what tools you're going to use or not?

And then we'll Yeah, building governance.

one of the first things that I usually recommend when building strategy, it's customers, I support, is a data platform.

Because we're doing type analytics, these aren't your silo based applications, like ERP, HR, et cetera.

You want a central resource that is especially built into, to see data, to process it, and then to deliver it to these users are on the farm.

And most large organizations have this, in some shape or fashion, but a lot of them now are kind of being challenged with these new data types.

And tried to do this for near real-time decision making.

Most of them and ah are not capable of doing that, and they need re-engineered.

Now, the financial services industry is probably the industry that buys the most, what we call alternative data.

It's a term that is used for data That's not in house it's external, but also kind of unconventional. It doesn't necessarily have to be a normal financial data.

So if you take a look at that graphic there, some of the data like satellite imagery.

Why would a bank or hedge fund want to buy satellite images?

And what some are doing, some tech trends are doing is they're buying it. So they they look at companies and they they track on activity.

They can tell activity by thermal readings by number of cars in the parking lots, and they use that as input.

two for soc evaluations.

So, those are things could say.

How much weight does that have?

Um, they put, they would argue that it's enough that they want to pay for it and use it as an input force. Or evaluation.

Just wanted to spend a little bit on the Collective Intelligence investing. It's an emerging area.

So if you're a Star Trek fan like me, the concept behind you, this is the collective, most better, than the individual.

So rather than you taking your stock, so you going to, a fund that's managed by 1 or 2 people, these photos are managed by a thousand people, and they all kind of share and collaborate what they think about, certain start, certain funds. And then, they break out recommendations into multiple risk classes, in which which class you fallen.

That's one they recommend to you.

And they claim that they beat the market, and any of the traditional funds that are managed, by just 1 or 2 people.

So, a lot of interesting things here.

The data that's being looked at, examined in mind, the Moses of social Media.

And, lagged far behind it is data from payments providers, a little more traditional there.

And investor management is the biggest group of financial services that are, aren't paying for so on to data.

On the right there that you see is that this market is expected to grow at 4%.

So, lots of focus here was determined by financial services.

Now, whether it's alternative data or internal data, things and vet, regarding the data, as practices, you have to cheat and say, you want to make sure you get quality data.

You want to try to eliminate silos, be open, have a platform to new sources of data, drive toward near real-time.

That's a hard one, not easy.

And build into it.

Privacy, security and ethics don't do this as an afterthought.

So there's lots of regulations some today, some new ones coming that will ask you, are you using this in an ethical fashion, or you're respecting citizens privacy?

Um, spend a lot of time on this, go lots of tools to accelerate. I have questions on this, feel free to ask me, but I just want to focus on was sworn to the robot.

And you don't have to be a rocket scientist to use this tool.

This will actually say, What algorithm do you want to use, join, use deep learning?

Do you want to use simple regression analysis, and you just check the data science algorithm, feed it the data, and it will tell you that say, hey, based on, But you asked me, here's my prediction for this, it's so simple to use it, it's not funny.

So, again, for those citizen data scientists, if you aspire to do things like that, there are tools out there.

And, again, as you get more, too, there's a data. Lots of tools out there to help accelerate.

If you have questions, I can ask them to answer them later, but don't want to spend too much time.

But these are the popular use cases where you want to do that data platform.

I'm doing analytics, supporting analytics on the edge from Edge devices.

Fitbit device. Do analytics, right there. As you will, more processing power, will do what's called ... data, stored in either data warehouses or data lakes for unstructured data.

And, again, yeah, machine learning or deep learning.

This is how you produce a rhythms.

All of these need a good data platform at the backend.

Hmm, hmm.

And, again, a lot of conversations about where to place this, lot of companies, a lot of banks, so, some banks are saying hey, we're doing cloud first, Tom, OK, great, hat's off to a lot of that is driven by time to market.

And I believe, to compete more effectively, with syntax.

However, I've seen a lot of folks stumble, Get Challenge Cherif because one Bank just put 4 to 5 petabytes of data up there in the cloud.

There is no cloud technologies ready to, to turn, to work on that day, in a timely fashion and then add the latency of it, and try and move to a real-time, It cannot be done with the current physical world tools that we have today.

So I believe what they are going to do, at some point, is re-evaluate and go take a business approach first. Let's look at the characteristics of the data.

Does this make sense?

Where are the costs long term?

Yeah, it's up there fast, but, you know, we're gonna visually break the bank, actually keep it up there.

Um, OK, last slide.

Well, actually, no. I didn't have a couple more slides after this. This is the last kind of presentation slide.

Um, the best practices here, regardless of what Division Marion, It's useful to have a very data driven culture.

It's almost, kind of, a pre work, if you're going to go, especially into AI, any kind of, analytics here, and have a strategy, Define the architecture first. before you dive in. No matter how compelling use cases, you can buy lots of applications out there off the shelf.

But, before you do that, work with the IT folks, try to define this architecture going forward. What, why we want an error here.

And, for each firm out there, this AI journey.

Now, my last last chart in the appendix talk about it.

I mentioned these references.

We'll be there.

Excellent.

Partner tools that you can look at there.

All right, here's my slide.

Yeah, so.

Instead of bull or bear market, as data scientists, you get a lot of data.

It's tough to avoid bias.

What does this tell you?

How would you interpret this data and if you're trying to create something?

So, um, proof that cursed.

I think I'll end my presentation for now and we can open up for Q and A It's an amoeba. It's not a polar bear. It's an amoeba. No, no. That's not what you ask. Oh, no. I'm sorry.

OK, I'm kidding, I love that the morphing of that, OK.

So, Bill, I am totally passionate and interested in your subject. I'm nowhere near as knowledgeable so, therefore I can be in the token question answer. And I've got a couple of questions that people have sent to me.

But I wanna, I wanna think in terms first of I think it was Clayton Christianson who wrote the book the innovator's dilemma Richardson who wrote that.

And the big takeaway that I remember from Clayton Christianson was these existing companies, whatever they're doing right wherever they're making are doing, are always under the threat of creative destruction.

That somebody comes along with an innovation and they just don't have.

you know, think of Kodak and film. There's a thousand examples, right?

They don't, they don't make the leap to the new thing for a variety of different reasons.

And you hear people like my son, who's in his twenties.

And he's in consulting, talk about, you know, all the Banks are just going to be put, to shame, that, are going to be wiped out by all these fintech companies. And you spoke to that. So if you're in a traditional financial services business, and that can be anyone that could be insurance, or it could be asset management, it could be investment banking.

What would you what would you like them to think about so that they can not be destroyed by the innovator's dilemma?

What, what can they do?

Yeah, that's that's a really good question and not easy to answer.

No.

The innovator's dilemma also points out a lot of times, the biggest kind of challenge thrips is really themselves the organizational structure.

They, they can have this new thing that they're developing really wipeout, you know, 80% of their organization which is really depends on.

So, how do you have this?

And some companies have had this, manage this dichotomy.

And there are books just written on just the manage at this two teams and how you can introduce this over time.

Even though I'm know that this one is kind of your own, uh, the breadbasket you need to depend on and everybody, all your staff, is really cocked on that.

And there are some best practices there. There are some best practices.

one try not to demonize that group.

That is mean pulling in 80, 90% of why you're there in the first place, Recognize that business is changing and how some companies do it is, is the organizational challenge is so difficult to devote. They have to partner.

A lot of the traditional banks are partnering with fintech to address certain markets mm because they find it's just too hard.

Um, perhaps they lack the political will to really kinda detonate, their existing org structure to, to, to make this work.

Maybe you really have to do drastic things here.

Um, and so, if you're not going to partner, you're going to do individually.

Um, you really have to, kind of, and there are kind of best practices of how you introduce this new technology.

And then, at what time do you introduce kind of more traditional measurements, like the other lines of businesses have today.

So, um, some are, are doing that now.

Yeah, It's a that is a big challenge.

No, I think that's a great answer.

I think about Walter Isaacson's book on Steve Jobs when he wrote that kind of a rushed biography empty jobs, because Steve Jobs realizes he realized he was dying.

And in Isaacson's book, he talks about the way Steve Jobs talked about the Apple II at when the Mac was being made, and he really treated the apple to people like garbage.

He called them the past and the legacy and there are a bunch of buttons.

And Wozny Act Steve Wozniak, co-founder of Apple, is saying, Steve. These people are paying the bills. They're the only reason you have money to develop a Mac. You can't be an *******.

And that sorry. That's exactly what he says, I'm quoting him.

And of course, Steve was famous for, you know, capturing the big a word.

And so I think, it just seems like that's, that's a lesson we, we want to avoid. Yes.

Yes, yes.

No, I mean, it is, so, why it is so challenging?

on a micro level, when I talk about this on a personal level, and I think sometimes it's easier to either understand something at the micro level, or at the totally gigantic macro level, in order to get to the real challenge, which is in the middle, and on the micro level, the leader walks into the room and says to six people.

We want the six of you to design a new process and use new technology. You can only needs three of you.

Was?

Successful at project is, I have a slide I use where, you know, the two guys are in a firing Squad, with their hands by there, and they're both looking at each other, going, Is it going to be you? Is it gonna be me?

And, I think that's a version of the same thing, right?

But, if you walk into the room, and you say to the six people, We have really work for nine people.

The only way you guys are going to be able to do this is, you're gonna have to figure out how to do this six person thing with three. So, we can put together three on the other part, right? Because there's no other way to solve this. So come up with a solution, and we'll figure out which three of you are gonna kinda keep doing what you've been doing, because you like it, and you're good at it. And which of the three, if you're gonna go off and do this now, I've never seen the team given that challenge.

That doesn't get all excited, right?

Because it's a hope and its future. And I mean, again, that's the micro level. That's the six person challenge.

But we're talking about go on to Credit Suisse, or Citibank, or Wells Fargo, or Fidelity, right, and saying, If you're not careful, you're going to have your lunch handed to you.

No. Yeah. That's a fascinating subject. You're brought somebody else I wanted to ask you about in this presentation.

Now, this is a political hot topic, but it shouldn't be, and it's a question of bias.

Because I spent years in financial services and credit risk managers are, by definition, looking for bias, right? They're looking for people who won't pay their bills.

They don't want to lend money to people who won't pay their bills, right. But they do it in a different, the way they go about doing that is, they look at trends In the last time we learned X amount of money to just kinda company. They didn't pay it back. So we're not spending any money those people anymore. And so how are you?

I mean, the good credit, you know, enough that if the organization is racist or something, that's a whole other story, but that's really what risk people are all about, is trying to figure out how not to do stupid deals. And they're looking for in the bias.

But sometimes they get it wrong.

How do you think AI can help get it right, and not have biases that are not warranted or justified?

So, one of the good things about AI, although it's a challenge, is, it can consume and process lots of data, and, unfortunately, humans, you know, a lot of people like to make decisions easy.

And if you take a look at banks, credit companies, when they do loans, um, and they take a look at people's scores, there are companies right now, uh, that were categorized by race, um, know, their credit scores or their credit worthiness, right?

They will split it up Caucasians to Hispanics, African Americans, Asians, and that that that is that is data that's produced today.

Well, aside from that being sounding horrible, is there any basis in that?

Mean, you know, that's a terrible thing to do if you're just doing it by race, but Why are white guys, like you likely to flake out on our credit card loans? Is that what you're telling me?

Some would argue, um, that, let's say within, you know, a few percentage, that, that, on average, it's, it can actually categorize your success rate.

It's actually within reason of doing that.

And she just, too, just to clarify, the highest rating were Asians with credit cards.

But what what AI can do for you, is let's get App, what we would call real contributors, Like like, let's say, years of work.

Um, years of education, number of defaults. This is data.

You can take the model and explain it to somebody.

Regulations say these are the reasons why we start dispersing.

These are the reasons why not.

And that's where hopefully I can help.

And that's where I think the model should go.

And I think I'm hoping the fintech's they get that and they build an allowance more like that.

It's lazy, people who kind of go, hey, let's, we have this data.

You all have data on their personal backgrounds and stuff, we can create this.

And right now, I guarantee you can find data that's characterized by rates on credit worthiness. No, I wonder if, going forward.

So, let's say, in 1950, the racial disparity in the way people pay their bills back was whatever, you know, you made some term chart, right? Well, guess what? The education levels, their work experience.

All this stuff is getting better for everybody.

So, you know, if you have a backward looking credit score, that may have been, you know, closer and closer to a time when groups of people didn't have opportunities, well, maybe they need to be start to do predictive.

Yeah, but those people have changing, they're getting, they're, they're better at whatever they're doing. So, let's start thinking about what kind of credit risk they're going to be in today, and as opposed to 10 years ago.

Anyway, yeah, it's fascinating that you can, you can at least get these indicators and then try to make decisions, You know, with what to do with them.

So I have one other question for you. Well, I just realized we're out of time. OK, I can't ask you another question.

That was, well, I gotta tell you, that was fantastic.

I'm going to go back and watch this in detail and make myself notes, which I don't have anything to write on right now. I'll make myself notes on your farm or your your previous presentation. I'll speak on behalf of anybody who watches this now and who watches this in the future.

If you don't appreciate how clear and how useful your presentation was, you should, because it was a great presentation.

I really enjoyed. it was just you and me. Well, I'd say, why didn't you bring coffee? But if it was just, you and me, I would say it was a great, great presentation. Thank you very much. So, I think we're gonna wrap it up for day to day. I feel you can. You can check out if you want, and I'll do a little bit of a wrap for tomorrow.

So, anybody who's here, us toward the end of this, I hope you didn't just check off and run off to lunch right now. We'll talk a little bit about what's going to happen tomorrow, we'll do a little warmup for what that session is going to be. These are three very interesting days.

So we're gonna start off today with a with a, an executive from Danske Bank. And I'll tell you later why that's a warm spot in my heart.

I lived in Denmark for about four years before I move back to the states, scatter scaled, Agile with global teams, and virtual working, fantastic ideas of what's going to be going on, and what's happening right now with virtual work and doing development.

How to develop a cost optimization ecosystem with TransUnion.

So another incredibly important point with cost driving so many things. Next generation digital financial services from William Genitive Easy from Huawei.

And then, we'll wrap the day with Oscar Stark, who I recently listened to give a great presentation.

Making rain the digital transformation journey of an African born and bred multinational ensure. So, if nothing else, those these three days, or a Bower potpourri of interesting topics, that will help you in your journey in the Amazon, your journey up to Mount Everest, your journey through the Gobi desert.

Or the metaphor on your head business transformation in your real life and real world. Thank you very much for joining us. I'm Chris Hodges. I'll be the chair tomorrow as well. I look forward to seeing your questions. I can't see your all your faces, seeing your questions tomorrow, and engaging with these interesting topics that we have coming up for tomorrow.

Thank you. Take care of me tomorrow. Bye bye.

About the Author

Bill Wong,
AI and Data Analytics Leader,
Dell Technologies.

Bill Wong is currently the AI and Data Analytics Practice Leader for Dell Technologies responsible for supporting Canadian commercial businesses and public institutions. He spends his time developing AI strategies for firms, providing guidance on how to accelerate AI development, worked with universities in advancing AI research, and publishes a quarterly AI newsletter focused on AI adoption in Canada.

Previously, Bill has held roles an enterprise architect in consulting and AI and analytics roles in development, and product management during his tenure at Dell, Microsoft, Oracle, and IBM. He also published numerous books and whitepapers on advanced analytics and data management and often speaks at conferences.