Session Information:

Data Science-driven Digital Transformation: Learning from Data-driven Organizations

Leading companies that have deployed Process Mining successfully have learned how to leverage data to drive business outcomes. Companies that are not data-driven have seen their data-driven competitors seize market share and make inroads into their customer base over the course of the past decade.

For several years now, investments in data and AI initiatives have been ranked by executives as the technology that will enable them to compete better in the marketplace. However,  despite this acknowledgment and investments, many organizations struggle to drive value from their AI and data analytics investments.

• By 2022, 90% of corporate strategies will explicitly mention information as a critical enterprise asset and analytics as an essential competency.
• By 2023, data literacy will become an explicit and necessary driver of business value, demonstrated by its formal inclusion in over 80% of data and analytics strategies and change management programs.

This session will review some of the major challenges and highlight the best practices and lessons learned from data-driven organizations.

Session Transcript:

From the beautiful city of Toronto in Canada, he has been a speaker in several of our conferences, because he is a true cross industry leader when it comes to artificial intelligence, and its applications and data science in general.

So, it's a real honor to have Bill Wong with us, Bill, in his current role. He is the AI and data analytics practice leader, working for Dell Technologies, are responsible for Canada. He has been directly involved with a number of public and private institutions to help them define their artificial intelligence strategy, and deploy a platform to accelerate business insights.

He has authored a number of books on advanced analytics, including a book focused on bio informatics, quarterly, AI, newsletter, and a contributing author with university researchers, for a paper that's currently under reveal on how data science algorithms can accelerate drug discovery, Bill. It's always a pleasure and a gift to have you here with us, sharing your expertise, and on behalf of our global community, we're very thankful for your time.

Thank you, Joseph. Saying, Thank you for inviting me. Hi folks. Welcome. Good morning, good afternoon, good evening, wherever you are.

And what I'm going to do is take you through this presentation.

I just felt, that talk about the, this merge is leveraging of process science, and data science.

So I find it really exciting area, it's still kinda new, which makes itself very interesting area to spend time with customers.

OK, so here's the agenda.

Give you a quick overview on digital transformation efforts. Most of the studies that I'll be sharing with you have been done by Dell, or sponsored by Dell.

Then we'll talk about process mining, and I'll give you a primer on what's behind the curtain, on how process mining leverages data science.

And then talk about lessons learn from data centric organizations.

And I know everybody's well versed on the process side, but there are a lot of parallels, that, and experience we can draw from, looking at the data side.

First step.

A couple of things about before we start, you see here on the screen here, these countries, some of the small countries like US And China.

United Kingdom, are well known for AI and, and, and and forth. Canada's actually didn't acknowledge as a worldwide leader.

And why is that?

Um, on the right-hand side here, what you see is we're the very first country to announce and to fund a national AI strategy.

So we have a number of centers across Canada helping companies, startups get into the AI world.

And some of the names you may have heard, if you follow those feel alive, Geoffrey, Hinton, Yoshua, ....

cool all cut their teeth in Canada and major research, major products have been spawned from these Canadian researchers.

So we're big in AI.

And like most things, you know, we we do keep a low profile, but don't be surprised, as you see in the comment here, by the Economist.

And a lot of things are the research that you're looking at. Canadian authors.

This is a survey that Dell does every two years and so this is the third iteration that I'm reviewing here.

Not surprisingly, digital transformation, we've seen a lot of efforts accelerated to the pandemic.

Everybody home, everybody being digitized.

And when you take a look at the top five programs here, not surprisingly, cybersecurity is so very, very important, even without the pandemic working remotely.

But we'll come to step four, using data in completely new ways.

More and more people are looking at data, and how to leverage that data and process mining as part of that, to drive competitive advantage.

Now, while driving for competitive edge is a good thing, this is not an easy task.
4:54
And here's the top three barriers.

I should mention that this survey that we did is a worldwide survey, 4200 C level executives, across North America, Europe, Asia, South America, as well.

So in getting access to data, one of the top concerns is, privacy security.

That's always something to be respected and should be designed into applications.

Budget was, that was a barrier for some people, I kind of have a philosophical view about budget.

I see sometimes that, really, that's enough.

Uh, really, it's leadership if you don't have somebody leading and trying to make data more democratize, and budget gets in the way.

But number three, time and time again, we see this as one of the top barriers to analysis is the ability to actually extract that data. There are a lot of technical challenges and business challenges as well that we'll talk about.

Now, try and become more data driven.

And as you do this, trust me, this will help you model business processes better, but it's not an easy thing.

Most organizations, if they've been around for 10, 20 years, have multiple legacy systems, which are incredible, dif, difficult to change, um.

And here, some motivations here. There's a quote here, that that's said often is that the companies in this, in the industry, they're more data driven.

Typically, our more profitable stock price is higher.

Any case in point, it's difficult to become this.

The whole thing is, is my technical challenge.

But a cultural challenge, again, the opportunity to talk law C level executives.

And I don't know if there's anybody from the construction industry, but my guess out there is.

No, that's, I've spoke to one C level executive, and basically, he told me, No, how we make decisions here.

It's more of a, kind of like, essentially, yeah, I've got feel. They have a good sense of what's going on all the time.

And this, this company was over 10 billion, and they, they use analytics, basically, just for simpler reporting. And that's it.

So, if you have that kind of perspective, mm at the sea level, very difficult to get any kind of data centric, kind of initiatives off the ground. that culture is just simply not there, to support it.

So, from a process based perspective, here are some of the major processes that you see in any company.

Finance, ERP, CRM, a chart.

All companies have these things.

Now, when it comes to leveraging data, as you can see, most everything is siloed here.

And a lot of the legacy applications, so they're, they're locked in these silos.

Now, when we've tried to leverage data from an enterprise perspective, and let's say you have processes that span across lines of businesses, very difficult to get at that data.

So, that is a key challenge that that remains.

This survey, and I'll share some of the results here throughout the presentation, is something that was released this year, that we asked Foerster to do for us.

What they did was surveyed a number of local institutions, and they asked, All right.

How do you consider yourself in terms of your daily readiness for digital transformation?

And if you take the date enthusiast, ..., Cindy, Technicians, that's 88% of the people. there are not the champions.

Meaning that they're being held back today, because the culture is not there, the skills are there, or the technologies are not there, so, uh, plenty of opportunity to, to improve care.

So I thought I kinda compare, love the process based technologies and AI and then take you behind the curtains about what we call data science, and I've worked in the past, I've been, when I was at IBM, I worked in development and while I was on the AI side, I would talk to the developers who are doing BPM and RPA.

And we kinda compare notes, how we do things. And so I came up with this kind of comparison.

The thing is, I would say, we're still in the infancy and I don't look at these technologies, as, you know, BPM versus RPA versus AI.

These can be very complementary, and I'm starting to see RPA applications leverage AI.

And while RPA is really kind of the mimicking of human actions, tass, think of AI, we're trying to mimic human intelligence decision making.

Now, as I was building this presentation, I wanted to kinda show how complimentary data sciences was process. And so I found this slide.

I've thought about this is one researcher who kinda put it together, a kind of saint perspective, and then I did some research on this person, and it's reference down below, It's in this second book.

I didn't realize it, the vendor S, he's considered the godfather of process mining.

So if you have a chance to read the book, it's great. It does NP positions process.

Mining is the link between process science and data science and I'm going to take you to the data side side.

So why elbow height.

If you survey sea level executives, you'll find that analytics, AI are right up there as the top areas, investment, or what are we going to do to differentiate our services and offerings.

So that's why there's a lot of attention, and currently AI is the fastest growing data center workload out there.

Another reason for the Focus and AI is there isn't one industry that remains kinda untouched, although maybe constructional will be one of the most industries.

I'm sure there are kind of leading edge construction firms that are looking at AI to, let's say, improve supply chain.

I spent a lot of time in health care, but starting to spend a lot time in financial services, They're probably one of the leaders when it comes to using new technology, Excuse me, but every industry has an opportunity to leverage AI, and these are just some of the use cases.

Uh, one thing I should also mention about AI is, there's also been a lot of hype, some companies, which I'm not gonna call out, have said like this, this thing can do anything.

And when it comes to investments, a lot of people whoa, rightly or wrongly, say we'll, you know, we'll get this approved because it's AI.

Let's not focus on these other kinda traditional technologies.

So let me be the first to say that not everything, not every business problem, not every optimization problem, is looking for an AI solution.

That's one of the first things I asked, you really need AI.

Can you use conventional analytics?

Is it something that is available to existing enhancements to your, your BPM, or RPA tools? So, I do not advocate that.

You know, you must use AI, but there are definitely things that are well suited.

So, you may or may not have seen this, but a lot of these terms are also used in process mining, as well.

If you take a look at the evolution of Analytics, from the bottom left-hand corner, we have what we call descriptive analytics, where it's really looking into the past and take a look at access, their time horizon shares in the past.

So what happened, just simple reporting, diagnostic, or why did it happen? So this gets into kind of discovery type programs.

And here a lot of process mining, also making this analogy of discovering no business processes. Why?

Well, I think that is kind of unique to AI.

Although it's starting to see some of this in process, my is that predictive analytics is trying to predicts who historical data, what is likely to happen.

And I do see this in process finding as well, is this term of prescriptive analytics, given the information we have, what should we do?

So, I know, I can't see people out there or comments out there, so I'm going to have to guess, or second, guess what you might think.

But, uh, hopefully, some of you have seen this movie called Moneyball and if you haven't, I urge you to see it. It's a fun movie.

It talks about the conceptions that people have in trying to manage a baseball team, and so here's the scenario.

Were the bottom of the knife bases loaded?

two out?

Who would you send to pinch it?

Now, baseball.

I suppose all sports are well known for keeping a lot of statistics.

and data science is science for trying to identify those variables that might affect performance or predictors here.

So common statistics that were captured and still captured today over the years are things like batting average homeruns runs batted in.

So one might think that you might want Jose to be your lead matter there, because he's got the highest value.

Or perhaps you want Austin, because he's got the most runs batted in.

Now if you saw the movie Moneyball, you'll know what the answers to this, and you'll have many people arguing based on the statistics, they have to support their argue.

What they found over time is that they really weren't capturing, are looking at the right statistics.

The statistics they should have been looking on is the on base percentage is how often that person gets onto base. Because that person gets onto base. Eventually, what happens is more runs are scored because of that.

And and the, you use this concept so that if you have, let's say, a babe Ruth type player on your team, who has, you know, an incredible average probably on base percentage, it's very difficult and very difficult to afford another bay routes.

And so what they did was they decompose the player to say, what is person's on base percentage?

And if it's zero point four zero two, like like mine is, what you could do is decompose and say let's get two players put together and gives us that zero point four zero two.

And you might even have money left over and that's what the Oakland A's found out.

And so that slide on the right is they made it to the World Series.

But that little bar shows that they had one of the lowest payrolls in all major leagues compared to the Yankees.

So typically, you know, the more you pay, the more you win but they were able to prove by looking at this predict this, this on base percentage, 80 completely revolutionized the game here.

And to this day, if if you are a baseball fan, if you've caught the World Series last year and illicit again the Dodgers versus the Tampa Bay rates. Tampa Bay Rays has the third lowest payroll.

We get a lot of people who don't even go to the Kings. And and the Dodgers had the second highest payroll.

And, and second, the only two to the Yankees.

And so if you you saw the World Series, then you know probably where I'm going with this game one.

The Dodgers won game to Tampa Bay one, game three, duggins, got ahead to games.

And then Tampa Bay tied it up again to those fees game to game five was won by the Dodgers Game. *** was a pill game. They had to tie it up to push it to a game seven.

They put their their best picture, southpaw, and he started striking people out like like crazy, under 100 pitches.

And then, in the I think it's a six hitting, they took out their star pitcher.

And they put in another picture, and in six pitches later, they lost the game. They were leading 1 0.

Yeah.

Everybody next day was wondering, Why on Earth did you pull out your star pitcher?

He wasn't even, it was way under 100 pitches.

Um, and what why they had such momentum.

And he, the coach never really said why he released. He just said, well, I thought he did enough.

But the folks of color analytics notice this. They were at the top of the batting lineup.

The guy who's about to bat was the is going to be his fourth time facing the picture.

And the trend is the statistics show that our bad are facing picture For the fourth time.

It's batting really improve.

So the concern was that, Hey, you struck them out three times, but the fourth He's going to hit.

And they took out the picture. And that's the theory.

And I think it is the correct theory that a lot of the baseball writers said.

And the headlines read that, know, the geeks rule baseball now, So it's out there. It's fundamentally changed chain game. And it's not just the data side. It's the process side, as well.

So what makes AI a little different here?

So, there's some terms of, I'll introduce here.

So, again, for AI, we're trying to mimic human intelligence here.

Machine learning is a set of algorithms at our disposal, to use deep learning as a special kind of algorithm.

It's trying to mimic how the brain works with neurons and you have multiple levels.

So, that is kind of, unique about AI, and on the bottom right hand corner, think of it as your traditional programs, you feed it data, and a program, have the computer, process it in an output.

In machine learning, what you're really doing is the output is a prediction program, feeding data.

You tell it what you'd like it to predict, and the computer figures out the algorithms, and that's your output. It's a model into producing these predictions.

What's also a little different about here, and, again, this is behind the scenes as the compute platform for this, uh, on the far left-hand side, we have traditional compute, CPU platforms, well serviced by Intel and AMD.

Then this newer firm called Nvidia, used to be well-known for screens, you know, high power gaming screens.

Graphical processing units are, are well suited because they perform vector operations really well. And if you have a background in linear algebra, deep learning algorithms are basically just an events in your programming.

So they do this very fast and in parallel, FPGAs, fuel programmable gate arrays are specialized processors you can code at the firmware level, and there are a couple of offerings out there.

I see this, usually, in a couple of industries, automotive, retail, but on the right-hand side here, a sip processors event, or such a sick, these are optimized processors that they do deep learning really well, and so, in a Tesla car, you have this four to do computer vision, etcetera.

So, these are the chips where all the startup capital is going, and they're building these into IOT devices, and you're seeing them also in new kinds of servers.

So if you're following this field, you may have heard of a chip manufacturer called ....

So they're attempting to challenge and video to develop a server, a shift, to do deep learning really well.

Application specific Integrated circuits.

That's it, OK, so when you go to IT, and let's say you're trying to do process mining today, it can be.

Let's say, a simple operation if it's just one application you're looking at.

So on the far left-hand side, you know, you have ERP data, or logs, event logs, et cetera, here.

That's the data that you're really trying to do with process mine.

This, though, if you're an IT organization, is you want to build a platform that can really host for all kinds of analytics here.

So the big picture is that you want all types of logs, et cetera, to be able the ingested.

And then what happens is that data gets curated in the middle here Where, where you have things like data warehouses, you try to consolidate it, you clean the data.

Then, to the right of that, typically, what the best practices to build a consumption.

So, meaning a data platform that's optimized for performance to service all the users on the right.

So mm best practice is to build, some people call this, a data lake data platform, that's something that is optimized for analysts.

So a couple of examples just to differentiate to kind of what AI can do here.

So all the credit card companies, you know, Visa, Amex, mastercard have AI algorithms for fraud detection.

They take a look at historical data and try to find patterns of what the bad people will do once they have credit card.

So Probably you've experienced this like myself is sometimes it stereo.

That's one of the, I guess, most popular items if a stereo pops up as something you paid with a credit card.

I've had multiple times, Eric. Erica Recall is that really Did you really buy a new stereo? Yes.

Um, more sophisticated algorithms that I've seen, and I had a person in one of my seminars, come to me and say, Hey, I had them ask me, was this you who use this? Because you want to try to avoid false positives.

And what this person did was presented in the capital Ottawa and he used the credit card to buy something here, It's just a snap.

And then, he was driving home to Montreal.

And midway, you stopped by to fill up with gas, and it got rejected.

Call it the credit card, company, So what gives?

said, I was, they used it earlier this morning, and now I hear affiliate gas. And it's not working.

And they said, Well, according to our records, yes, you did use it this morning, but where you show them gas, there's no physical way that you really should be there, at that time.

And he laughed, and he said, OK, you got me.

So he was speeding, fortunately, too, too reckless.

But, yeah.

He was speeding quite a bit, and if you've ever been to Montreal, you'll know that that's not an uncommon thing.

So, they can get very sophisticated, these algorithms.

Um, and the slide I had before, some of the use cases that you'll see for AI or deep learning is computer vision.

And so, too, improve supply chain.

This little robot that you see learns the layout of a warehouse and can make deliverance.

And they have versions of this with arm, so you can pick up things in the warehouse place, sit down yourself, delivered to another part. And these are perfect applications that can get integrated with RPA, patients, something that is very repetitive.

So, lots of opportunities for these technologies to work in concert.

So, now, some of the challenges that we have, and there are plenty well, while everybody says this is great, this is not an easy task.

So, here, we see some paradoxes are quite out. So, on one hand, people are saying, we need more data.

We need more data.

and, on the other hand, here, you know, even more people will say, We can't access, we can't handle the data.

And part of the problem a big contributor is, when you get into analytics, you need people who are educated about out and so here, 61% here in our survey with Forrester said that many are held back because of insufficient training.

7% in this mission skills.

And yet, while they say this is all important, just very small percentage, companies are actually doing anything about it.

to hire data experts.

And you don't have to be.

data scientists to, to do these things.

What's emerging in the industry are, what I call citizen data scientists, so people who are well versed in an application and what data it uses, and, you know, they're, they're somewhat technical, and all you do is you provide them the tools and more and more, you will see application development in this area, where you don't have to code and, you know, Rob Python.

It'll be components that you'll just assemble put together.

A best practice is always to build a strategy, and if you are an Enterprise Architect, you'll recognize the steps of defining your vision, You know, assessing your environment, creating what we call, what, the art of the hospitals, and then defining a roadmap.

How to get there. Kind of low hanging fruit to get there first, and the business case.

And then, finally, governance, and this is where the data privacy data security, should be designed from the start.

And here, for your reference, these data science best, practices, and challenges all apply to process mine, as well.

You want to try to get data that's clean.

You want to avoid silos, nice, incorporate enhance data by leveraging other sources of data.

The more advanced applications I see is applications now, that can leverage near real-time data.

And, again, privacy, security ethics, I believe, is very important.

So, this is my last slide.

Um, and, as I mentioned before, what we can do to prepare, and the most important thing, is to have a data driven culture, All right?

It's not just a technology problem, and we really have to be, you know, fact driven organization and the leading companies, You know, the internet companies, they are very data driven.

Take a look at Uber, Netflix, et cetera.

They are hungry to process data to better understand their customers, avoid data silos.

And, again, successful technology architectures have a focus on both the data and the processes involved here.

And so it's not one without the other. You really need both of them. Both of them have to be looked at.

That's it for my presentation here.

Fantastic, Bill. It's always a masterclass when it comes to artificial intelligence and digital technologies and the data sciences. So, we had questions coming from the audience on a different.

Different areas that you've covered during your presentation here. And, as I scan and look for, the overall themes, continue to pose the questions to, to Bill, here and now relay as many questions as possible in the time that we have allotted.

The first question has to do with building the right skill sets in the organization.

You know, data science is, is a big field and when you wait, as you're building your tribe, as you mention farewell, You need to work on culture. You need to become more data driven in the organization.

And then as you start building infrastructure to support that journey, what kind of skills you should be looking at in your data scientists? Maybe technical skills, but maybe some of the soft skills as well. Well, how do you screen for data scientists in this day and age?

It's interesting question because we have we were involved in hiring Adele.

So we do look at certain skills and there's no, let us say uniform template.

I'll start with the hard skills first. Yes, It's good to have.

Uh.

In data science, that's why I didn't take statistics. I didn't consider, you know, mathematics. I have a mathematics background. So.

And for the Rod deep learning, it's good to have a math background, but you don't necessarily have to have a computer background.

I've seen institutions where they take a physics major and they enhance and train him, and he became their data scientists.

I've seen organizations where they say, look, I don't ever want to code, but I think I can become a citizen data scientists. I know what the important data is. I know how to use it.

I need somebody else just to get me that data.

So, yeah, there's no, let's say, one profile for that.

Um, but for the soft skills, I think that's equally important.

A lot of people, I've seen, at the beginning of their hire, university and master's, PHDs, Right out of school, giving them 300 asked authors and say, Hey, develop this AI project.

And some succeeded, but many did not.

They didn't have soft skills like how to communicate, how to work in teams.

Um, I think those are equally, sometimes equally, if not more important, because you can't do this alone.

It's, it's too big and too complex for one person. Do it And you don't want to do at home.

You want to the information knowledge to be democratized over as many people.

So yes, you want somebody who's good technically, but we also want good soft skills in a team environment.

It, for sure. That that's well well said.

That it does require interdisciplinary skills and the hard skills for sure and the right soft skills, collaborative leadership skills to go with that. I will, I will, let's, let's shift the conversation a bit towards Moneyball and First of all, I'm deeply hurt that I didn't make the first choice on the selection of the team. But you made the right decision.

You you found a cheaper player with a better on base percentage and that should be the person you go for.

And that, now it's interesting, I've actually worked with Billy Beane before, you know, the, the guy behind limitation at the Oakland A's and you should talk to Billy Beane about that.

He will tell you that the data science is very important.

But it was not the most important thing, because the, when he first implemented moneyball in the Oakland ace, it didn't work because the data was right, but the manager would not follow other datasets, which is the so-called process owner. So that he did not have a buying from the process owner. He had the key influencers in the organization, which was a specific player in the clubhouse, home-made way more money than everybody else was not on board. So, even though he was told that he should pay to play in a certain way, based on the model, he didn't, he realized, found very smart ways of sabotaging it. And because he was a key influencer in the clubhouse, the other players were looking at him for his behavior instead of the directions from the data.

So, this shows the complexity of things, even if you have the right data set.

And even though the data science behind it is sound, the implementation doesn't work if you don't get people to accept it, right?

So, highlights, again, the importance of what you just said about those.

Those, we need to understand how change happens in organizations, and bring those skills to data science.

Now, it's hard because we're asking, you know, people should be really good at a hard stuff and they're really good at the soft stuff and it's not easy to find those people, right? I mean, I think there is, there is a there's a real gap in the marketplace right now for, for the combination of the skill sets. Is the, is that how you see it as well? That it's a real challenge. Sure. Find those people and then attract them.

Yes.

First, thank you for sharing that story about Billy Beane and when I watch that movie too, I think the thing I get most out of it is the cultural battle maybe had to get this and I just want to re-iterate why it's so important to have that data culture.

You can have the science there in front of them.

by the less people buy in and buy in is a soft skill, a soft sell.

And it's unfortunate and we see a lot of people fail because he can't get the right people on board.

You know, this is such an important conversation, because there is this perception, especially from, you know, I'm an engineer at heart. I understand this. I'm, you know, I'm a physics engineer guy. So, and we all are hardwired in school and college specifically to think that data is, is, right? You know, science is sound, and everybody will kinda follow along, but they don't. And I think ... is a great example of that. You can, you know, don't bother me with the facts I've made up, My mind is a real thing in organizations. And so, I'm glad you brought this up, in the context of process, mining and artificial intelligence, because these are some of the real hurdles of getting things implemented.

Yeah, it is hard mm, but it can be transformative. The advantage of process space.

Like BPM, RPA, it's very easy to say, listen enough, I cut out two processes here, make this more efficient, I'm going to save you X percent, Very simple, and then the science isn't that difficult to understand.

So, it's easy to get buy in.

Now, if I told you well, customers might feel better than they might buy, or the quality of your decisions will be better.

That is not easy, too, Costs justify, and a lot of people come up with another argument, say, well, I don't think I think this will happen.

So, while the benefits can be enormous, the challenge of adopting is, it's equally challenging care.

So, and tier two, the back to the skills.

I find organizations, too, are, they put out these ads for people, but when they come to the organization, they're only paid to do one task. well.

Fortunately, What we find is that people who can disrupt and change things, they look beyond the sign that they're in, and unfortunately, people don't get paid for, that people are compensated for that. And a lot of times, they're penalized when I say, hey, you're going out of your swim lane here.

So to be able to have a comprehensive perspective, I think it's absolutely necessary. But there's not a lot of rules are asked to do that.

one exception, perhaps, might be an Enterprise Architect.

who's asked to look at, the business, Asked to look at technology, to ask, looking at an application perspective.

But outside of that, like even the data, people said, Hey, we just want you to get the data. We got the process folks here, they're gonna look at that.

But if you want to optimize this and really transform how the business operates, you need somebody for people to look across and companies don't pay for that. We don't train people like that. You kind of have to learn this on the fly.

Bill, what you have just shared, I'm going to go back and I'm going to record this segment over and over again on the playback, because that is gold for the less experienced ones watching this.

They may not have understood that that before you just said, but if you have been around for a few decades and seen organizations actual work, and how successful implementation of technology and culture change takes place, what you have just said is what the count of that is? What great enduring organizations do in a very structured way. They'll call the things like 20% time when you're expected to get out of your a little bubble and go across the organization and find something that that actual oil collaboratively create disproportional value.

And then, it's structurally doing something to counter this force that you just mentioned that is so critical. Wow, we can talk so much just about that.

I want to ask you a technical question related to artificial intelligence right now to wrap up this segment and just finished with that with that tone. Well, maybe technical, maybe a bit more strategic, as well. AI has been hyped a lot, and when you're getting behind the scenes and you're using it, sometimes people are using multivariate regression analysis and they think it's a, I guess I'll come on people know what does it cognitive thinking that's going on there. So, there has been a lot of hype on AI and Ms. Attributions related to AI.

Where are we at today?

Is it I mean, we know lots of use cases and lots of applications that happen our day to day even of our phones, but do you think that is still being hyped too much? Or, it's This is a time we have gotten to the place where we just need to go exponential in the applications because it's ready for prime time.

I think it's still being overhyped. I think the technology right now is still in its infancy.

So, if you have an AI application that does something well, a lot of times the learnings of that cannot be transferred.

So, let's take an example where Watson, from IBM did really well at playing Jeopardy when they tried to take that technology and use it in healthcare.

They had lots of challenges and many would simply, they failed miserably there.

So, but there are plenty of opportunities And we're at that point again where the science is there, but the people adopting it is the challenge.

So there have been studies in one in Toronto where they, they may come up to radiologists and they say, we want you to rate this recommendation.

What they don't know is, sometimes the recommendation, they were all done by computer, but we told them, when one group that was done by human and other group, we told them, it was done by AI and they always ranked the human better, even if they were incorrect.

So, there are a lot of things still going on and a lot of miss perceptions of what AI is still.

So, we're still in early days.

Science is still exciting. There are still breakthroughs that are happening. We still got a long ways to go. After I still get questions on, you know, is Skynet coming or anything like that now?

Many years before that, We're still on the technical perspective, still, in the early days, adoption lots of opportunity, but, again, it's the the culture, the human aspects that we're having challenges with, right now.

That's excellent, Bill, your presentations are always tremendous because you share market research. That is so insightful. I'm, I'm reflecting the comments and feedback and the, and the praise from our audience that has been shared throughout this conversation. And I wanna let you know that, on behalf of our global community, were deeply thankful for you to take the time to share expertise, your insights. Your real, practical applications, of how AI can create value today, and how we're going to continue to create value in the years to come. Thank you so much for, for being with us.

Thanks very much. Have a good day, people.

Thank you, Bill.

Ladies and gentlemen, that was bill log, artificial intelligence and data analytics, the leader at Dell, and the real cross industry leader, when it comes to artificial intelligence and digital technologist.

Always a pleasure. If you play back this session, this session, pay careful attention just some of the data that Bill has share with us. There's some real great insights built into that dataset that he shared with us, so we are going to be wrapping up this session.

I encourage you to check out on the on LinkedIn what speakers are saying, what participants are saying about the conference questions that they have, that they have brought up?

You can go under LinkedIn under my name, shows up areas, and you'll see there, my last feature post, is the post about this conference, and that's where everybody's making comments about what's going on here. So if you're deriving value from this session, make sure you go there. Thank our sponsors for allowing this to take place. Thanks to our speakers, to bring in their thought leadership, and being so generous with their skills, and time, and sharing that with us. We appreciate that. We're going to wrap up here for a break and when we come back, we're gonna come up with shoe industry leaders coming directly from Europe. We're talking about Mark McGregor who is the author of Performance and Business Coach on Major Digital Transformations and Sophia Passover, who is the founder and president of Stereo Logic. And they're gonna talk to us about all that Glitters is not gold. Why broadening your approach to process mining is the key to long-term success.

Sessions with Mark are always intriguing and interesting. So I hope to see you at the top of the hour.

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.