Subscribe
Share
Share
Embed
This is the Executive Connect Podcast - a show for the new generation of leaders. Join us as we discover unconventional leadership strategies not traditionally associated with executive roles. Our guests include upper-level C-Suite executives charting new ways to grow their organizations, successful entrepreneurs changing the way the world does business, and experts and thought leaders from fields outside of Corporate America that can bring new insights into leadership, prosperity, and personal growth - all while connecting on a human level. No one has all the answers - but by building a community of open-minded and engaged leaders we hope to give you the tools you need to help you find your own path to success.
AI is just a tool to achieve an objective. So the objective is whatever you want it to be. Do you want it to save you money? Do you want it to save you time? So just as Microsoft Excel was a tool, just as the calculator is a tool, AI is increasingly going to be seen as just a tool.
Raj:The real question is, what do you want that objective to be? What problem do you wanna solve? And then it's really about finding the right application machine learning or computer vision or generative AI to make that happen for you, to help you achieve that objective.
Intro:To the Executive Connect podcast, a show for the new generation of leaders. Join Melissa Arschog as she speaks to a wide variety of guests that bring new insights into leadership, prosperity, and personal growth. While no one has all the answers, by building a community of open minded and engaged leaders, we hope to give you the tools you need to help you find your own path to success.
Melissa:Welcome to the Executive Connect Podcast. I'm so excited to have Raj here today to talk with us about AI strategy and education. Raj is the director of applied AI research at Blue Artificial Intelligence, a consulting firm that specialize in AI strategy, projects, and education. He also teaches university courses covering business applications of AI and emerging technology. Welcome, Raj.
Raj:Hey, Melissa. Great to be here. Happy you could have me on.
Melissa:I'm gonna jump right in and just ask you from your perspective, can you share with us what initially drew you into the AI consulting space?
Raj:Yeah. So my background and my path into AI and AI consulting is not really standard by by most people's, understanding. So I'm not a programmer. I don't I'm not coming from a tech background. I actually got my start in the insurance industry out in Asia.
Raj:I work for a French multinational insurer called AXA. And I was sitting out in Hong Kong, and our office was running digital transformation, IT transformation programs for our businesses in Hong Kong and Southeast Asia. So I actually did that for about 7 years, and I got a really, front row seat into how not only the insurance industry, but also financial services in general was moving from paper based in person ways of doing business to doing business over the Internet, website based, mobile based, and really collecting a lot more data from customers and the business in digital format and taking action on that data using basically intelligent IT tools. So I did that for about, like I said, 7 years, then I went off to business school. I went to the University of Hong Kong to do my MBA there.
Raj:And while I was there, like every MBA students, I had a choice to make. What do I wanna do with my future? And I had it all planned out in my head. I thought, okay. Insurance is great, but I wanna pivot into finance.
Raj:So I'd done my CFA, got my CFA credentials, passed all 3 exams. I thought I'll move into investment management. But as it turned out quite, fortuitously, I ended up taking a course at the University of Hong Kong called AI for Business Leaders. And when I took that course, it wasn't just an introduction to AI from a technical standpoint. It also covered how it's applicable for enterprises.
Raj:And something really clicked in my head because I thought, hey. In my past life, I was doing this digital transformation, IT projects kinda work. And now that I'm learning about what AI can do for the enterprise, I really felt that digital transformation had laid the groundwork for using machine learning to automate business decisions or at least make more intelligent business decisions based on data. So I thought, hey. That's exactly where I wanna be, and I ended up getting an internship for a short amount of time and just stuck with the company I'm with.
Raj:And the company I'm with is called Blue Artificial Intelligence. We're based out of Canada. We have offices in Canada and Hong Kong and just do business all over the world.
Melissa:I love it. Isn't that how life happens? Sometimes we just pivot on a dime. If you would've told me 10 years ago, I'd be in the casino gaming industry working in cybersecurity, I would have laughed to you. Here we are today.
Melissa:Right? It's funny how we pivot and things that interest us and our passions become things that we explore. You know, as it regard as it pertains to AI, what type of business problems is your organization solving right now?
Raj:Yeah. So that's a great question. And what my company does is we do a mix of AI education at the executive level, and we also do do strategy consulting that really builds upon the AI education and awareness. So in terms of the business problems that we solve, we really have found out that any company in any industry, forgetting about AI, specifically, they have to solve 3 business problems or do business in 3 ways to optimize these 3 metrics. And the metrics are getting business insights.
Raj:What do your customers want? How is the market evolving? How is our company performing? And then do we have to pivot? So business insights.
Raj:The second is how do we reduce risk? So if you're a bank, how do you reduce the risk of fraud and money laundering? If you're a supermarket, how do you reduce the risk of running out of inventory? So risk reduction, very important. And finally, there's cost reduction.
Raj:Not really firing all of our employees, but really finding ways to do more with the same amount of resources that a company has. So those are the act the 3 problems that we solve or we help our clients solve. And if I can give a couple of examples without really naming the clients, we work for factories that manufacture clothing and apparel, and we help them build a computer vision based defect detection system where you have a literally a mobile phone that's plugged into the assembly line taking, real time videos and saying, oh, this product is defective. Take it off the lot. This product is fine.
Raj:Send it on to the next point. Things like this just saved so much time for the factory itself and, like and it's just helped with quality control. It's a very simple tool. We didn't need a lot of hardware. It wasn't a high cost project, but it had real results.
Raj:And that's really on the risk reduction, cost reduction side. And if you look at the business insights, so one of the other projects that we've done a little bit of a while ago was working for a property manager that owned a bunch of commercial real estate, and they wanted to maximize the occupancy within their within their proper property. So getting more customers in and really finding a way to maximize occupancy in the place without increasing prices too much. And this wasn't really a machine learning kind of solution. It was more data science driven.
Raj:And we just found a way to say, okay. Here, based on the data that you have, customers coming in, going out, what kind of vacancies you've got, here are some ways you can maximize occupancy or increase occupancy without really increasing prices so much. So we just brainstorm a bunch of ways where you could use the input from statistical modeling or data science to give yourself inspiration on how to make, more impactful business decisions.
Melissa:That is interesting. One of the areas that I'm the most fascinated by in AI is the intersection between AI and cybersecurity and how companies can use AI tools for to help support cybersecurity. I know, you know, I I joke I joked recently on a panel I was on where back in the day, I used to get a lot of these phishing emails, and they spelled Microsoft wrong, or they'll click here, and the logo for Wells Fargo was wrong. And, virtually, overnight, all of this has become so hard to spot. The it's perfectly written with a perfectly colored logos, and it they're hard to spot.
Melissa:So I know, you know, the intersection between AI and cybersecurity, I guess the question is, how does your organization or, help mitigate some of those risks now and maybe your insights as AI crosses into cybersecurity, AI tools?
Raj:Yeah. This is an area that every company is looking at, and I can speak to this from 2 standpoints, 1 as a consultant. And, also, I teach university courses on AI when I'm not working. So when we go into the classroom, how do we talk about cybersecurity for younger business leaders? I can kinda share both stories.
Raj:I guess one of the things we tell our corporate clients is pretty much exactly what you mentioned. Like, people are using ChatGPT to write really high quality phishing emails, where if you're an overworked employee and you get this email, you're like, this looks legit. I have to provide my credentials because I keep asking for them because I need to reset or I need to, like, upgrade my assets. Let me just do it. So what's the way around this?
Raj:I mean, in a way, it's the solution is less to do with cybersecurity and just, like, more making humans trust emails less, like, be more critical and really look through it. And, I mean, at the old company I worked, even before all this AI stuff, we had a saying that said security is everybody's responsibility. Not that we're kind of putting their blame onto the employee that clicks on the email link, but just company wide education saying this is what a phishing email could look like. These are the things that our IT departments or human resources department would never ask you for. So just raising awareness to kind of, put up a shield saying, okay.
Raj:I got this email. It looks legit. But then 1 or 2 things don't make sense because my own company has told me that we would never do something like this. So that really starts with educating the workforce on what a reasonable email request could be. So that's maybe the most effective short term way to counteract this AI written fraud fraudulent emails.
Raj:But, I was just I just wrapped up teaching a university course over in Hong Kong. It called big data and finance. And I I mentioned it was in Hong Kong because literally when I was teaching the course, there was a really high profile, like, deepfake scam where, there's this Hong Kong company, Financial Services. They didn't name the company, obviously, but a fairly entry level finance clerk got really scammed by a deepfake video that was pretending to be the CFO of the company and a few other high ranking employees that made a request to, say, transfer $25,000,000 into an offshore account. And, of course, this employee, he's not, dumb.
Raj:Like, he was smart. He's been doing this, for a few years. He was like, okay. I this doesn't make sense. So the scammer said, let's get on a video call so you can see our faces.
Raj:And the faces were completely deep faked. The the audio was deep faked because all these people were relatively famous. There were clips of them speaking at public events, and these scammers just took those clips, trained their models to basically create a filter, fatal filter to appear as the CFO or this finance director, and they got this for employee to transfer $25,000,000. And that money is essentially gone. I hope that guy is doing okay.
Raj:So I guess how do you mitigate these risks? I mean, it's very hard. It just starts with education. But then there's another angle here, which maybe you can appreciate from a cybersecurity standpoint. It's just how do you detect when a hacker is potentially in your network?
Raj:And, really, the answer to that, which companies are working on, is just building neural networks to kind of understand what does normal network activity look like within our servers and train a model with examples of normal network activity and malicious network activity. So the model learns over time. It gets better and better the more examples it sees. So when there is somebody that's potentially intruding on the network, the system could automatically flag it saying, okay. Investigate this.
Raj:This doesn't look right. So we can go from the very low tech ways of just making people more aware to spot a scam where and also building a machine learning model to automatically detect very hard to find malicious network activity to kind of not solve, but at least mitigate some of these cybersecurity concerns.
Melissa:Absolutely. Well said. I agree. I know, you know, if if if you follow me, I'm always traveling somewhere, this side of the globe, that side of the globe, and learning people's behaviors is key. You know?
Melissa:And I don't usually work at 2 o'clock in the morning. Now I might if I'm in another country and it's a different time zone. And so I think you just understanding your people, when they work, how they communicate. And, really, I hate to use the word, but common sense. Right?
Melissa:Do do we normally send, you know, this large of amount, $25,000,000 at midnight? Do we normally is there level as levels of approvals that need to happen? So I think you it's education. People are our best line of defense, and it's also stopping and thinking about, you know, what you're being sent is, should I get a second opinion? I had it happen to me recently where they used my social media profile, how I communicate online, and they went and targeted somebody else that knew me well.
Melissa:And I saw the email and the communication. I'm like, wow. That sounds exactly like me, exactly something I would say, and it was so similar that it's almost like they put one of my social media post in chat g p t or something and rewrote it, and it sounded exactly like me and something I would do. And so they're getting to be so, so good. And so, like you mentioned, mitigating the risks and protecting our employees because, like you said, that that employee probably lost their job, and, unfortunately, it's it's hard.
Melissa:It's hard now. Right? And so and then kind of pivoting when I think of generative AI, you know, how how do you think generative AI is transforming and changing other industries?
Raj:Yeah. So I mean, we can talk so much about this. We can look at the impact on creative industries, such as your writers, visual artists, designers. You can also look at corporate white collar work. Like, how is it changing these industries?
Raj:I mean, my general piece is so I use a lot of generative AI tools. I use Chat ET. I run these models open source models on my own computer. I use Midjourney, all the video generation stuff. And my general takeaway, regardless of which industry you're in, is that we can anybody out there can use generative AI more as a thought partner, kind of to augment their own thinking, their own work, and speed up the time it takes to produce an output, whether that whether that output is a blog post or a marketing post for your company.
Raj:And it's really about augmenting content creators that are already effective. It's not going to automate everything that we produce. Like, honestly, if I look at chat CPT generated writing, most of it is really bad. I would not read it. Whereas if I read a blog post or an article from, let's say, a journalist or a blogger that I follow, that, like, resonates with me in a different way.
Raj:So AI is not good at doing this yet, but the same author or blogger can say, okay. Hey, Chat GPT. Help me brainstorm ideas for this topic, or here's something I've written. Am I missing anything? So I like to see generative AI used more as an augmentation tool rather than a straight out automation tool.
Raj:Yeah.
Melissa:I like that idea. I might have to use that, Raj, is asking ChatGPT or Grammarly or some of these other tools to am I missing anything? That's a real that's a good tip. I'm gonna use it. Pivoting a little bit from an ethics standpoint on AI from an ethical standpoint?
Melissa:How do you think companies should be paying attention to ethic?
Raj:Yeah. So, I mean, I I keep seeing this beam online where I believe it's from IBM from a management training seminar they had really in the eighties or even before. They said that a a computer cannot be held accountable for a business decision. Therefore, they should never be allowed to make business decisions. So I I might be paraphrasing it and taking creative license.
Raj:I believe it was IBM. If if not, it was somebody else in that, in that league. But that really applies to how we use algorithms for decision making today. And because an algorithm can't really be held accountable for discriminating against one of your customers or making a decision that loses the company money, it's really important to have some level of human in the loop operations. Now from an ethics standpoint, every company or industry is going to take a different approach.
Raj:So, for example, if you look at banking versus advertising, if in banking, if you if you if your algorithm discriminates against a customer that wants a loan, or a group of customers based on their background or their income category, that's generally bad. Like, the reputational risk is high, and the fallout from generally being unethical or discriminatory is pretty bad. Now so banking, financial services, really customer based, client based industries, they're going to really look at, putting in robust systems where these AI systems that are making decisions or making recommendations, that the models are explainable in some way, where you can go in and audit the model and say, why did you make this recommendation to make this loan to a customer or to, let's say, do this kind of operational activity. And generally, there's probably gonna be a lot of human oversight where, in a perfect world, there's going to be employees that have the right to veto what the model recommends, saying we're going to do something entirely different because based on our expertise, and our understanding of the business and the customer, we think that what the model's recommendation does not make sense.
Raj:So that's kind of one way to mitigate the ethical damage from relying on models too much. And when I teach about AI and we we get asked the question, what will AI do to careers in the future? One of the things I say is it may actually give rise to something called a corporate AI ethicist, ethicist, where you have a department or a person or a team sitting in a company where their job is to really look at all these algorithmic decisions that we're doing. Are they actually serving the needs of our customers? Are they making our customers' lives better instead of just focusing on top line revenue or or bottom line growth?
Raj:So, really, the the ethicist in this sense is not really it's really not a machine learning engineer, but somebody that understands a business well enough and also understands technology well enough to say, okay. This is where the bottles are operating in an acceptable way, or this is how they can be improved so that they meet the needs of customers better without violating, just normal social or ethical or business norms. So that's something I really see accelerating as more companies become more data driven, more model driven when it comes to making decisions?
Melissa:Yeah. I'd probably go out on a limb and say a lot of companies everybody's using AI now. I don't think I know any company that's not using some form of AI tool in their sales or marketing or finance department. I I would you say that most companies now are using AI tools?
Raj:Yeah. Absolutely. They're using AI driven decision making, absent, for sure. Because even if you're looking at, a tool like HubSpot that you people use to manage their customer customer resource management systems, marketing systems, there's a lot of AI tools built into HubSpot. So even if a company themselves is not using AI as their core business is is not something they're designing, They're definitely relying on third party tools that are using machine learning in some way.
Raj:And if we look at more data driven industries, they absolutely had to be using more AI for operational decisions. So even if you're like a supermarket, you may think, oh, supermarket, it's a brick and mortar business. People go in. But then every supermarket that's of a particular size, they collect a lot of data on who goes in, what do they buy, and when do they buy it. So they have a pretty good idea of the customer's behavior, customer preferences.
Raj:And they're kinda using that understanding of their customers to really do better demand forecasting. So they figure out, okay, what kind of products are in demand at certain times of the year or certain times of the week? Which products have seasonal demand? Is there any weird spikes in demand happening, like, during the pandemic when everybody bought toilet paper for some reason that I still don't understand. And and even at this point, you can the supermarket, which is traditionally seen as a old fashioned, not data driven business, they can plug in all of these data inputs into an AI system or a machine learning tool that kinda makes predictions about what should be stopped, what should we have in stock, what do our customers want, which product should we market to our customers so that we get the biggest bang for our advertising dollars.
Raj:And then these models can plug in seamlessly into, inventory ordering systems, procurement systems to automatically replenish stock before you actually stock out. So to your point, every business is using AI in some way, either consciously in ways that we don't think about, like supermarkets, or unconsciously when a small business is using HubSpot, using Midjourney, ChatGPT just to make their daily life easier in some way.
Melissa:Yeah. And I see a lot of buzz right now on sustainability and climate change. And, you know, I know all over the world now, everybody's having weird weather. Places that didn't snow is now hot, and the hot places are now raining and snowing. So, you know, I I know it's become really popular on online now, AI and climate change.
Melissa:How do you any thoughts on on AI as it pertains to sustainability and climate change?
Raj:Yeah. We could talk about this all day, and, the issue really starts with there's no real accepted understanding of what climate change is. And the truth is many people just don't take it seriously enough to invest into it. So I actually was speaking at a climate change conference a couple months ago. And when I went in, I was like, alright.
Raj:So we have we have metrics to track climate change, whether it's the 1.5 degrees warming temperature, like sea levels rising and so on. But most people that doesn't resonate with most people. You just have jokers saying, oh, if it gets a little bit warmer, I would like that because they're very cold here in Minnesota or so or wherever, or in Toronto where I am. So my real thing when it comes to how AI can make climate change more manageable or counteract the effects of climate change is just reduce pollution and waste. Like, how can you kind of do this?
Raj:And you show them a picture of New Delhi where it's smoggy all the time. Show them a picture of Beijing where it's smoggy and it's polluted. That really gets with people. And then they ask the question, how can AI make this better? So the real impact that AI can have, not just to reduce emissions, but also to save companies money, is optimizing energy use, not using less energy because over time, we've used more and more energy.
Raj:This is gonna continue. There's no such thing as a low energy prosperous economy. So the real question is, how do we not waste energy so that it we're just not burning coal or burning gas for no reason? And AI can really help here. So if you're an office building or if you're a data center, you're spending a lot of money to cool the building, to keep the lights on, keep the air conditioning in the data center running so your servers don't overheat.
Raj:So what machine learning models can do is they can take data on temperatures, how many people are in the building, what's the air quality in the building, and it can kind of train the model to understand at what times of the day should the air conditioner be on, what times of the day should lights be on, how can we modulate energy usage so that things are comfortable for the people or these servers in the building without just using too much energy. And one of the people that did this very early on was Google, really back in 2016 where they found a way to use machine learning to automatically modulate the air conditioning in their data centers. And they managed to reduce their electricity bill by 40%. And you and that's a huge number for a data center. You can imagine the similar effect that it could have on a commercial office space.
Raj:And you're not just saving money here. You're also reducing your energies and your energy usage, and that's just great for most countries that are still reliant on natural gas, fossil fuel for their energy needs. And if we look at any city in the world, it's not residential units that are using a lot of electricity. Commercial commercial real estate manufacturing, they use a lot more energy. So if we could find ways to optimize that or modulate that, that in my view can have the quickest impact on mitigating the worst effects of not just climate change, but increased emissions.
Melissa:Yeah. That's 40% is a humongous number for humongous number. I didn't realize that was that high. So living in Austin, you know, everything's AI right now. You can't ever it's the big buzz.
Melissa:Everybody's anticipating the next big thing. There's tons of startups here, and there's more moving every day. And so I'm curious from your perspective being in the education space, what are you the most excited about as it pertains to AI and, you know, as we continue to develop, you know, AI and generative AI?
Raj:Yeah. So what I'm the most excited about is really this overlap between AI and augmented reality. And I should explain this a little bit because what is the most cutting edge ARVR tool we have today? That's probably the app Apple Vision Pro or the MediQuest. They're really bulky.
Raj:They're hard to use. You look like a kind of a dork when you walk through the streets wearing them. So but then the proof of concept there where you can overlay a layer of intelligence really over the real world, that has some real impacts. So in going to the future as these models get better or as these pieces of hardware get better, we could really come back to the era of smart glasses that overlays information about the real world in front of us. So your point about, whether you're living in Austin, whether you're living in, Toronto or somewhere in Asia, people are just doing normal people things.
Raj:And how can AI kind of integrate into that length experience and make it better? So literally a couple of examples. If you're wearing smart glasses, that gives you information about the real world. You got a camera on the glass. It's looking looking outward.
Raj:It's giving you feedback. It could be as simple as, hey. I'm in Austin. We're at summer. We're we're barbecuing.
Raj:Home my glasses are telling me just take the steak off the grill now before it starts burning. They're just little little tips like that. Or let's say you're on a first date. You are dressed to impress, and then you get feedback from your, smart glasses saying, oh, based on, your date's body, body language, she's really not into this right now. So maybe time to tell a funny story, have her think have her think that you're interesting and a fun person.
Raj:Just little tidbits like this that seem futuristic, and, and you it may be hard to believe now, but that mix of augmented reality and just reality, just our physical reality, if AI can somehow feed us information to kind of have us navigate life in a more seamless way, I I see a lot of potential there. And I'm pretty sure in the next 3, 4 years, nobody's gonna be wearing these bulky headsets. It's probably gonna be smaller, leaner. It may even be other devices that you can clip onto your ear that's even less obtrusive, but it's still giving you feedback on the real world in some way. So that's personally what I'm really excited about.
Melissa:Yeah. And when you said that, it made me think of the old cell phone, like the big bulky old cell phone with the top high antennas, and now, you know, our iPhones are like our computers now, our walking computers. And it was a really quick time frame that I went from, you know, this to this. So I'm also interested to see how it develops. Coming from the education space, just kind of in closing here, some key takeaways for the listeners that you wanna leave as it pertains to AI from your perspective, maybe top 3.
Raj:Yeah. So first and foremost, AI is just a tool to achieve an objective. So the objective is whatever you want it to be. Do you want it to save you money? Do you want it to save you time?
Raj:So just as Microsoft Excel was a tool, just as a calculator is a tool, AI is increasingly going to be seen as just a tool. So from an education standpoint, whether I'm in a classroom or talking to executives, the real question is what do you want that objective to be? What problem do you wanna solve? What opportunity do you wanna take to take by the horns, really? And then it's really about finding the right application machine learning or computer vision or generative AI to make that happen for you, to help you achieve that objective.
Raj:And I guess the last thing I'll say is, like, AI is evolving so quickly. There's there's a fear among people that they're kinda getting left behind or there's too much information to to, take on board. So my general advice there is just try using some of these AI tools. You're not that far behind. You're you're actually very, very early.
Raj:Most people have not used ChatGPT. It's free to use. Just log in and use it. See what it can do for you. And if you like it, keep exploring it.
Raj:If it's not for you, that's totally fine. But then in the future, if we're in one of these jobs that are potentially at risk of AI doing more of it, I don't think AI is gonna replace most jobs. I do believe, however, that people that are good at using AI to do their jobs might get more job opportunities than people that are not good at using AI for that same job. So it's a never ending learning experience, and hopefully we have fun with it along
Melissa:the way. I love that. People have to tell the tools what to do in the AI tool. So it's a person that's telling them what to do. The tools aren't telling us what to do.
Melissa:So it's really great points, Raj. Thank you so much for your insights, your information, and your knowledge in this space, and connect with Raj, Yvonne Lincoln. Thank you for your time, and that's Executive Connect.
Intro:You've been listening to the Executive Connect podcast. If you have questions or ideas on how to bring leadership to the next level, email us at executiveconnectpodcast@gmail.com. And don't forget to subscribe so you can catch every new episode. Until next time.