EP 346: AI Hype vs. Reality - Insights from EY Consulting Leader

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life.

Is AI just hype, or is this the foundation of how we work? You know what? It's. It's something you see studies and you see talk happening on both sides, people saying, hey, there is no return on investment for generative AI initiatives at the company. And then there's other people kind of like myself that say, no, this is literally how you work in the future. So we're going to be tackling that big picture today, head on and talk about AI. Is it hype, or is this the reality? And we're being joined today by a guest from one of the leading consulting firms in the world. Ey. So I'm excited for today's conversation. All right, what's going on, y'? All? My name is Jordan Wilson and I'm the host of Everyday AI. And if you're new here, well, this show is for you. This is a daily livestream, podcast and free daily newsletter helping everyday people learn and leverage generative AI to grow their companies and to grow their careers. There's always so much going on in the world of AI, so this is your spot every single weekday where you can keep up with it and not get overwhelmed and get ahead. So if you are new, thanks for tuning in. Make sure to go to your everydayai.com sign up for that free daily newsletter where we will be recapping today's show and going over a lot more. And make sure while you're there to check out our thanks a million giveaway. All right, so before we get into today's topic, let's talk about what's going on in the world of AI news. So Google has unveiled a new version of its Gemini 1.5 flash and Pro updates. Google has rolled out significant updates to its Gemini AI models and including Gemini 1.5 Flash 8B and improved versions of Gemini 1.5 Flash and Pro as part of Google's push toward Gemini 2.0. So the new Gemini models are designed to handle long context and multimodal inputs, managing over 10 million tokens. Yes, I did not miss anything there. That's 10 million tokens. Wild. So Gemini 1.5 flash is highlighted as a top choice for developers, while the updated Pro variant is positioned as a replacement for early, earlier models. Starting September 3rd, Google will transition users to the new model and phase out older versions. Community feedback so far in the first couple of hours has been mixed. While the Gemini 1.5 flash has improved its Chatbot arena ranking significantly, some users are pointing to ongoing issues with coding tasks and repetitive outputs. All right, next Nvidia's anticipated earnings calls today is set to impact the tech sector sector as a whole. So Nvidia is poised to release its second quarter earnings today and the results are highly anticipated across the tech industry. Investors are closely monitoring these figures for insights into the future of the AI market. So Nvidia is expected to report adjusted earnings per share of $0.65 on revenue of $28.7 billion, marking a remarkable 139 increase in EPS and 113% rise in revenue compared to the same quarter last year. Wild so the companies, despite a reported delay in the company's rollout of their next generation Blackwell GPU chips, which is what powers essentially generative AI analysts, believes that these reported delays will not significantly affect the company's performance in the near term as existing production of hopper chips continues to improve. And Nvidia is obviously a leader in AI chip production and its performance sets the tone for investor confidence in the AI sector. All right, last but not least, OpenAI is reportedly not working on one, but two major updates at once. So we got this into our newsletter yesterday, but it happened after our podcast. But according to a report from the information, OpenAI is racing to introduce a new AI product called Strawberry. We've talked about it here on the show a lot, which promises to improve reasoning capabilities beyond what current models can achieve. So the new AI product is expected to tackle complex problems more effectively than existing models, potentially transforming how chatbots interact with users. But here's where we got some new news from the information so OpenAI is also working on another model called Orion, which will utilize synthetic data generated by Strawberry to enhance its performance and versatility, creating a new kind of relationship between the two products. Also right now, reportedly the Strawberry features are being tested right now for the federal government, and both models are part of a broader strategy to improve reasoning and problem solving skills in generative AI systems, addressing the growing demand for more sophisticated, sophisticated interactions in various applications. All right, a lot more on those news stories and more, so make sure to go to your everydayai.com all right, but let's get into today's topic. So I'm excited about this one we're going to be tackling Is AI just hype or is this the reality of how we work? All right, so I'm not on this alone. Please help me welcome our guest for today. Bring her on the show. There we go. There we go. We have Tracy Gusher, who is the America's AI, data and automation leader at ey. Tracy, thank you so much for joining the everyday AI show.

Yeah, thanks for having me, Jordan.

All right, I'm excited for this one. So, Tracy, can you just tell us a little bit about what your role is there at EY and maybe for those that aren't familiar even what EY is and what they do.

Yeah, yeah, happy to. So EY is one of the big four global audit, tax and consulting firms. So we have over 300,000 employees worldwide, operating in well over 100 countries. And I have the great pleasure of leading our AI and data practice for the Americas region. So thank Canada to the tip of South America and live in Philadelphia. So kind of in a little bit north of where half of our business is. But practice that I have the great pleasure of leading is AI engineers, data scientists, automation engineers, data engineers that help our clients utilize data to garner insights to automate processes and ultimately find ways to generate value for them.

And hey, as a reminder to our live stream audience, it's not every day you get a leader like Tracy on the show, so please get your questions in now if you do have them. So, Tracy, I want to dive in a little bit more on what your role actually entails, because I think the data side is something that so many people overlook. But if we can real quick, I kind of want to skip to the end and talk about. Is, is AI, is it hype? Or is this the reality of. Of how companies will be working in the future?

Yeah, well, you know, certainly there's some hype out there, but AI as a. As a topic area is not hype. The the reality is, is AI has been around since the 40s, at least in concept, and we've gone through something like three AI winters where investments slowed and interest and perceived value to come from AI waned. And we've had four cycles of waves of peaks of interest. And during each one of those, we've seen value. And what's different about this cycle is that for decades, those of us who have been working in AI have been talking about the democratization of AI. We've been talking about how do we get AI into the hands of everybody versus just data scientists and engineers behind the scenes. And that's what's different about this wave, is that AI is becoming part of all of our lives, both our personal lives and our professional lives. And it's being put into our hands in a very different way. We can actually use it to do things that prove fruitful for us, that, that prove valuable for us, versus just kind of cool, fun, you know, kinds of, kinds of trick tricks. And, and that's what's different about, about where we're at right now. And, and the reason that I say it's, it's not hype is that one of the things that generative AI has done is it's, it's released excitement about the promise of AI and it's created this influx of investment, not just in the tools, but in looking at where we can actually apply it. And what's that done is it's not only opened the door for generative AI to provide value into organizations, but it's reminded everybody of what classic AI can do in organizations. And the two together, you know, classic or narrow AI as well as generative AI, the value is there and organizations are finally starting to realize it.

You know, speaking of that, Tracy, organizations starting to realize this, Right? And you've talked about how, yeah, AI is not new. It's been around for many decades. How has the demand changed? Right, because I assume you've obviously been working with, you know, thousands of clients over the decades, you know, in deep learning, machine learning, and companies that have invested in traditional AI. But how has it changed now with the, you know, this generative AI wave in large language models? Like, what are you now hearing from, from your, you know, I'm guessing tens of thousands or hundreds of thousands of customers worldwide when it comes to AI and just what they want in general?

Yeah, well, I think that what they want is, is, is value. And for the organ, most of the organizations that we serve, ultimately what they are trying to drive is shareholder value. Right. So how are you driving top line growth? How are you improving the bottom line? And so what they're seeking to find is with this new technology, generative AI, where can I apply it in the business so that I am gaining that top and line, that value generation, generating value for our organization and for our shareholders. And that's coming in a lot of different parts and pieces depending on size of organization, where they are at in their maturity scale from an AI perspective, where they're at with in their maturity scale from a data perspective, and that they all kind of want the same things. But what, what those things are, is, is very different organization to organization.

Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on gen AI. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you, you can partner with us too. Just like some of the biggest companies in the world do. Go to your everyday AI.com partner to get in contact with our team or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen. So now I, I, I do want to rewind here a little bit and you know, talk about this, this divide almost between like, hey, is AI just, just hype or is it kind of the future of work? And we're going to be sharing about this in the newsletter, but you know, EY just came out with a new the AI Pulse survey and it said that 95 of companies surveyed are investing in AI. But you know, you see all these other surveys also, Tracy, that say maybe only 5 to 10% of companies are actually implementing it top to bottom, right, company wide. Why do you think there's, there's this huge gap between everyone saying, oh, AI is the most important thing versus those that are actually implementing it top to bottom?

Yeah, yeah. Varies from company to company, but there's a, there's a few, there's a few differences. So, so the first is that many organizations like, like you're saying, are experimenting with our, with AI in different parts of the business. Right? So let's see how we use it in product development, let's see how we use it in finance. Right? But they're not saying like let's transform the company using AI. That's, that's one bucket of companies. There's another bucket of companies that is saying this is, you know, this is the Satya Nadella moment. Right? Satya and Adela pulled in his, his product engineers three years ago and said, burn the ships. We are rewriting all of our products with generative AI as, as a piece of that roadmap. And there's a lot of organizations that are taking that approach and saying really like we need to rethink what we do, how we do it. What we are taking to market, how we are taking it to market, and there's a big difference between those organizations and why they're doing it. First is where they're at in the overall maturity curve from a technology perspective. If you are not already well into your cloud journey, if you are not already well into a significant, significant initiatives and programs to improve data, if you are not already a company that is used to agile innovation, then it is really difficult for you to now go back and say we're going to burn the ships. Right? So, so just those maturity scales are a big difference. And I, and I saw a question pop up here in, in the comments here that I think is germane to this neck. The next point, which is a lot of organizations are not going top to bottom because they're concerned about, about privacy, about security and about other types of risk areas. Right. So especially when we start looking at regulated businesses like life sciences, like certain parts of health care and beyond, there's, there's a lot of concern about not just the technical aspects of using AI in their business, but also the reputational concerns. Right. So what happens if I improperly use data? If data, the data I produce in a model is biased and then a poor medical decision is made because of that biased data? Right. What happens if I am creating different types of credit modeling and biased data or biased models are created? And by the way, just on the bias topic, we can't eliminate bias, so we should stop using those words. We can mitigate it, we can recognize it, we can't eliminate it. But those types of reputational concerns as well as risk and privacy concerns are definitely preventing a lot of organizations from really going big.

So. Tracy, 100% agree, right? It seems like even in the kind of quote unquote, earlier days of generative AI data has always been almost as mysterious as the black box of generative AI itself. Should companies be doing right, and I love what you said there, you can't eliminate bias in data, but you can mitigate it. But what should companies be doing to tackle this, you know, this big data, you know, mystery of data? Because I like, they have it coming from all places. You know, it used to be in this warehouse, it's in this data lake now. Like what should companies be doing to make sure that their data jives with the future of generative AI in their work?

Yeah, so the first is that so many organizations treat, treat data like a, the, the part of their business that they just don't want to deal with, that they, you know, it's kind of like that necessary evil in the organization. Like we know we need it and we know there's value there, but we really don't want to spend any money on it. And we don't really see the value out of the money we're spending. It's kind of this nebulous thing that it's intangible. You can't like a product on the shelf when you repackage it. You can see, look how pretty it is now, right? There's not like that, right? So the intangibility of it makes it difficult to really want to go big on spending huge dollars. But the mistake that a lot of organizations are making is I like to say they're playing whack a mole with their data, right? Like they, there's a project that we're working on, there's a new analysis we want to do and the data is really bad. So let's go fix that data. Like let's just go fix this data. And then they fix it on a project basis for this one specific use. And then they say hey look, our data is really good in this report or it's really good in this one model. But the deluge of data behind it and the repeatability of even that data being good is lost. And the organizations that are getting it right are treating, are treating data like a function that is as important as their finance function or their HR function, right? They are, they truly have a data function that proliferates into the business, right? Like think about your HR organizations and a lot of, in a lot of companies, right? Yes, there's a central HR org that is managing your benefits, that is managing your onboarding and the recruitment process is right, centralized. But the HR business partners that serve the business are out in the business, right? They are servicing those businesses. Data should be no different. You should have central data processes that dictate the way the organization should be managing, governing its data. But the actual hands on keyboard management of that data needs to be happening in the business. And that kind of central and spoke model of data management as a fully funded enterprise type of function just isn't happening in a lot of places. We look at finance and say we have to fund finance, right? We have to produce financial statements, but we don't look at data that way. And it's, and it's a huge mistake because ultimately your HR department, your finance department, your product department, your supply chain department runs on data. And so if you're not putting equal emphasis on your data department as you are, those Other functions, how are those other functions ever going to do what they need to do? And AI is no different. If you want to do AI anywhere in the business, if the data that you're going to run it on isn't high quality, isn't highly accessible and, and doesn't have the kind of governance on it that you know, other critical processes have, you're never going to be able to trust the AI.

You know, Tracy, I think that's such a good call out because, you know, I'd say maybe a year or two ago companies thought like, oh, if we use generative AI, that's a differentiator and it's not anymore. Right. I think that's the level playing field and you really have to have your data, you know, your data has to be in tune in order to actually differentiate yourself from your competitors. But I'm curious what is maybe a common thread and maybe this is whether how companies are using data correctly or just how companies are approaching generative AI. But what's a common thread that you're seeing? Whether it's across, you know, ey clients or just, you know, your own observations. But what are some of those common things that companies that are getting generative AI implementation? Right. What are those companies doing the same?

Yeah, so there's probably a few things. One is, you know, I'm going to sound like a broken record here, but they're, they're paying adequate attention to their data and they are, you know, we like, we like to say that you're, you're creating AI ready data. Right. And AI ready data doesn't mean just that your data's good for AI. It means that you've got good data that is good enough that it can be used for AI. And there's, there's a lot of things under the hood on that. Right? So, so that would be one, they're paying adequate attention to their data that. The second would be that they are thinking about AI as a truly innovative, market moving type of enablement versus hammer looking for a nail. And when I say hammer looking for a nail, a lot of organizations that I've been working with are saying like, hey, how can I use this in this function? How can I put some tools in the hands of the people in this function? So they're using generative AI in their everyday jobs and there's value in that, right? There's value in individual productivity, but that's not going to create shareholder value. Right. And the organizations that are getting it right are rethinking entire processes the organizations that are getting it right are saying, how do I create a new business, a new product line, a brand new revenue stream using generative AI as an enabler. So that would be number two. And then the third is they're not looking at generative AI as a single threaded tool to solve problems. The connectivity of using generative AI with applied AI with automation, with analytics supported by good data, that's the magic sauce. I like to say that if you break down an entire process and say, I'm going to rethink this entire process and you are rethinking it with the lens of how do I use all of those things, automation, generative AI, applied AI, analytics with good data, you're going to get huge value out of, out of that kind of initiative versus let's look at the procurement process and see where our procurement people can use AI in their everyday jobs. You're not going to get market making.

Kinds of, of value out of that market making value. I think that's, hey, everyone out there, write that one down. That's good. We're gonna, that's gonna be in the newsletter. So, so Tracy, one other thing and I, I like, I can't imagine because I literally spend hours every single day keeping up with generative AI. That's what we do here at everyday AI and I even struggle trying to keep up with the pace of innovation. How, like, how big of a, of a factor is that for this whole like AI? Is it hype or is it reality kind of scenario that we're playing out here? Just the pace of development and how fast the space is moving.

You know, Jordan, I got this question, I was on a panel a couple of months ago and I got this question and you know, the way that they had posed it was how do you, how do you keep up with it all? And my response was, you can't. And I mean, that sounds like a cop out, right? But, but, but, but, but, but in all honesty, if you think that you can keep up with all of it, you're fooling yourself because the pace is so fast. Yes, you can, you, you, you can keep pace, right? You can, you can, you can follow the majority of it, but you know, the, even some of the things that you were highlighting at the top of the show right now what OpenAI is doing with Strawberry and beyond. It was just a year ago that we were still kind of really grasping the last OpenAI release. And then Sora came out. We all went, oh my gosh, this is amazing. And now just a few Months later, we're looking at Strawberry releasing. And it's the pace of movement that is happening in this space is extraordinary. And, you know, all we can kind of do is run behind it and try to try to figure out how best to know where there is development happening that is going to really help us versus just some cool new things. And that's what I focus on. What I focus on is where are the developments happening that have real impact to those kinds of value areas versus cool new tool to use in my personal life. Right. That's the question in the lens that I put it on it every day.

And so many cool new tools and questions. But speaking of questions, maybe we have time for one or two here before we have to wrap. So, Tara, asking here, in your experience, what are your initial steps for identifying problem areas in a company's data and how do you determine the key metrics to, to measure.

Oh boy. So we could have a whole show on this. But so, so first and foremost is do you have, do you have data governance, data governance as an enterprise type of capability? And the reason I say that is because if you don't, that means you don't have proper data policies, you don't have governance processes that are followed, et cetera. So that is one just a key to understanding if your day, if you have, if you don't have those kinds of processes and policies in place, your data's bad. I don't even have to look at your data to know your data's bad. So that's one second is though, is that there's a lot of technology and processes out there to do things like analyze data quality, to do data profiling, to look at the lineage of data. All of those are measurements, right? So, so can I track the lineage of my data profiling and quality tools will look at what the quality of your data is in terms of comparing it against the standards of how that data should be, should sit in those systems, how many hops your data is taking between source system and where it's being consumed by an application or by a model. That lineage of the data is another measure. And I could keep going on and on and on and I'll bore your viewers to sleep, but there certainly is tons of different ways to both analyze it and measure it.

All right, I think we have time for one more here. So Fred's asking, how much work have you done with corporate boards regarding their responsibility for AI risk mitigation? And I'll add on top of that, how responsible should boards be when it comes to AI implementation.

Yeah. So what's interesting about the board view is we've actually been working with a lot of boards and the way that we have first engaged with a lot of boards on this topic is. Educate me. We actually have something we call our AI Masterclass and it's a class designed specifically for executives and boards. And it's one part kind of classroom facilitation and one part hands on keyboard. And you haven't really experienced the aha, wow moments of a board member until they've created their own GPT model. Right. And the, the, the, the awestruckness, the, the, the moments of. I had no idea it was really this powerful. Is, is pretty, is pretty extraordinary. And, and what that also does though is it opens the conversation for really cool new tool, really great opportunities. But now let's have a conversation about how you were going to be thinking about doing this responsibly. And so it's a, it's a glued topic, but we typically don't get to the, the risk components and the. How do we be thinking about using this responsibly with an organization at a board level until we've educated them, until they really understand what they're saying, dealing with.

So we, we could keep this conversation, Tracy, going on for hours. We can't keep you for that long though. But maybe what is your. As we wrap up here, what's your one most important takeaway? Whether it's it's companies who are already engaging with a consultancy firm like ey, or maybe it's smaller medium sized businesses that maybe that's not in their budget. But what is that one most important takeaway? When we look at this AI hype versus reality.

Can I have two?

You can. Yes. Make the rules here.

All right, so I'm gonna, I'm gonna double back on some of the things we've already talked about. The first is don't think of AI, not just generative AI, AI as a single productivity lever. Think about AI as a mechanism to reimagine processes, reimagine new revenue streams, reimagine your products. That would be one. And then the second would be, please, please, please don't forget about your data and how important it is. I tell the organizations I work with for every dollar you spend on AI, you're going to spend 20 to 40 on data. And if you're not kind of using that as a lever for how you are thinking about funding your AI projects, you're going to be willfully disappointed at the outcomes of the AI.

So many great takeaways. Wow. I'm I'm excited to go back and re listen to this conversation and write today's newsletter. But Tracy, thank you so much for taking time out of your day to join the Everyday AI show and help us all better understand this AI hype versus reality conundrum. I think you put it to rest. So thank you for joining the Everyday AI Show. We appreciate your time.

Yep, thanks so much Jordan. Have a great day.

And hey, as a reminder everyone, there's more. If this was helpful, please repost this to your network and go to your everyday AI.com Sign up for that free daily newsletter. Tracy just gave us. So much value we're going to be recapping it all there. So thanks for tuning in. Thanks for joining and we hope to see you back tomorrow and every day for more Everyday AI. Thanks y'.

All. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit youreverydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.