How to gain insights on navigating the generative AI revolution

In this episode of the Better Innovation podcast, Jeff Saviano joins Roger Park to delve into the intricacies of GenAI.

Related topics

The spotlight on generative AI (GenAI) is intense, dominating headlines and sparking novel experiments in many companies. Yet amidst the fervor, the technology’s emergence raises critical questions. Do leaders truly understand its implications for their businesses? What investment and deployment frameworks should they be considering?

Joining host Jeff Saviano is Roger Park, EY America’s Innovation Leader, renowned for guiding companies through disruptive change emanating from digital technologies. Jeff and Roger explore the complexities of GenAI, examining its origins, identifying investment opportunities and establishing the groundwork for effective deployment strategies. Whether you're a seasoned innovator or just dipping your toes into the world of emerging technologies, this episode promises to leave you inspired and equipped to embrace the future of GenAI.

Key takeaways:

Artificial Intelligence (AI) experimentation and deployment have surged in the business community.
Widespread adoption of GenAI is driven by user-friendly interfaces and open-source accessibility.
The shift to new AI platforms is intensifying, becoming integral and irreversible.
Organizations are moving from process optimization to comprehensive AI redesigns for transformative outcomes.
Worker upskilling in AI capabilities is prioritized to maximize productivity.
Hyper-personalized AI products raise ethical questions about transparency and data privacy.
Businesses must innovate and adapt strategies as AI evolves, considering ethical issues and regulatory developments.

For your convenience, full text transcript of this podcast is also available.

Intro

Meet the people behind today’s leading innovations – from the boardroom to the halls of government. Join Jeff Saviano, a global innovation leader at EY, to hear from the trailblazers reshaping our world. You’re listening to Better Innovation.

Jeff Saviano

Hello to my Better Innovation friends, Its Jeff. There is such a bright spotlighting on generative AI. It is so intense and it is only intensified and its been dominating headlines and sparking worldwide debates. How do you capture value, but also how to regulate this powerful technology. We are getting important questions from our client like , How can AI fit into their business? Or what kind of investment and deployment framework should they be thinking about for their company. Well you are in luck today to help answer these questions we have my friend and collegue Roger Park with us in the studio. Roger is EY Americas Innovation Leader and he is also a member of leader team at EY .In this role Roger is helping clients thrive in era of disruptive change. He is lazor focused on activating innovation and he embraces the emerging technologies and he is driving new solutions and business models, he is with us today to dissect Generative AI. How do we find investment opportunities and he gets such a rich point of view of the framework for businesses to affectively deploy AI. Here we go my friend Roger Park.
Roger Park, welcome to the show.

Roger Park

Hi, Jeff. Glad to be here.

Saviano

Roger, we finally did it. Feel like we have been working on this conversation for jeez! probably a couple of years now. We finally found a time to get together and have this important conversation. I appreciate that.

Park

I'm looking forward to it. I know. Couldn't be a better topic. I know we're going to be talking a lot about innovation, but also AI today. So, the timing ended up being pretty good Jeff.

Saviano

This is like the perfect discussion for my friend Roger and I. How many conversations have we had about innovation over the years and especially now at the intersection of AI. So, welcome not only to Better Innovation, but welcome to our special series on AI. And we're coming to a close and just another episode or two. But this is an important conversation. We've been looking forward to this, Roger, to have our audience hear your point of view. So, before diving into generative AI, why don't you talk about and provide an overview of your background and the roles that you've had at EY.

Park

Yeah. Sure. So, I'm currently the Chief Innovation Officer for EY in the Americas, and I lead our innovation business unit that operates across all of our service lines, our consulting business, our tax, our audit, our strategy business across all of our sectors and our regions. Think of it as our internal venture arm that's really charged with sourcing and incubating and scaling our best and most innovative ideas at EY. So, in a lot of ways, I feel like I have one of the best jobs at the firm. But before that I led our technology consulting business and our financial services practice, and I've been a partner at the firm for quite a while now, not as long as you. Jeff, but long enough.

Saviano

We've both been here forever, it seems like, and we've both been on our global innovation leader team together for several years. And you've been helping our AI innovation teams and maybe that's a place for us to start. Roger, as we're having so many conversations with our clients about innovation and the intersection with AI, but from an innovation perspective itself, maybe if you could start and talk a bit about what it means for a role like yours and innovation officers at our clients, how is AI helping to support innovation?

Park

Yeah. So, that's a good point. So, AI as you mentioned, Jeff, AI has been a significant portion of our innovation portfolio for years. New and emerging technology, fast moving, very disruptive, real impact on the business and significant value to be generated there. But in the last year or so, the activity and the proportion of our portfolio that is focused on AI has increased dramatically and I would say that that's been around the same time that ChatGPT was launched. So probably 80% of our portfolio, Jeff, is really focused on innovation in terms of new and early-stage ideas that we're funding and managing and sourcing. And also, I sit on our EY.ai steering committee, which we just launched a few months ago, and that manages our AI programs globally. So, I get good visibility into everything that we're doing across the regions, across service lines and all areas of AI. And I can tell you it's pretty exciting, very exciting stuff.

Saviano

It's hard to find solutions that are not influenced by AI and we saw that even before the explosion of generative AI. But if we focus on the innovation process to self and it's of course it's not always so linear right from idea generation on the front end, I love the phrasing, the front end of innovation. How do you source ideas and brainstorming and the generate within generative AI has been so helpful there. All the way through creating content for innovators from presentations and prototypes and conducting market research from an innovation process and from that strategic domain. Talk a bit about how you think generative AI is supporting and enabling the innovators.

Park

Yeah, well, first of all, I think the muscles and the capabilities that you flex, you know, getting reps in on innovation are the organizations that have done that, that have invested. I think it's really paying off now, when you talk about being immune to the disruption. So, things as basic as being able to rapidly test and learn with new technologies, being able to engage with ecosystems rapidly, you know, very very dynamic environments with new players and old players and all those dynamics changing very quickly, being able to safely innovate, understanding the guardrails that need to be put into place, understanding how you set up the sandboxes, etc., and then connecting to value, not just about doing proof of concepts with the technology or doing what would be more research and development on the technology, but also understanding the applications of those technologies into the business in ways that can provide value, in ways that are implementable. So, all the things that we do around innovation Jeff that are value adding to a large enterprise, how do you innovate at scale efficiently, safely and connect to value? All those capabilities that we all have and need to be innovative organizations? I think they're all getting exercised as we're, as companies are responding to the opportunity and sometimes the threat of AI and being able to move quickly in that space. So, I think those skill sets very much align. And even more so looking forward, everyone's looking at the pace of change in technology. But what we're seeing in the AI space, Jeff, is it's really quite incredible how fast the technology is moving, how quickly it's expanding and how fast we as organizations and large enterprise enterprises have to move in order to keep up with the change and respond to the change. So, all those things, I think, point back to an innovative culture, a platform for innovation, and then having good, strong processes around innovating safely and efficiently.

Saviano

That's really helpful Roger. I like how you phrased that. Let's get to the why. I want to get to the why around AI. Why do you feel AI became so hot in 2023? What changed in the landscape of AI investing that there was right around, as we're recording this conversation in December of 2023, that it was right about a year ago when we first started to see the significant hype that then only accelerated throughout 2023.

Park

I think it was several confounding factors Jeff. I think the headline event was the launch of ChatGPT 3.5, November 2022. Very exciting. I think it's still one of the fastest adoption curves of any new technology that's been out there and that really kind of put generative AI in the forefront socially and then in the discourse. But as you know, there were a lot of things that were happening behind the scenes that led to that breakthrough event. Traditionally, AI has been around for decades, but I would say that there were few signature breakthroughs, probably in AI technology that happened in the last few years that really opened the door to ChatGPT and then just the emergence of generative AI late last year and early this year, one being I think in 2017, DeepMind, the Google folks published their white paper on transformer technology and as you know, that's the T in GPT. But that was quite a breakthrough when it came to generative AI algorithms and models. That was a turning point on the neural network side. There have been more advancements since then, but there was a significant one. The other changes are really more around or related to the exponential growth of data and compute that really enables us to tap into the power of AI or deploy in a commercially viable way and consumerise it and all those things really came together in the last few years. And then when OpenAI made the decision to make ChatGPT 3.5 broadly available to the public, that I think definitely opened the floodgates.

Saviano

It's so interesting over the years on, let's face it, on multiple occasions there have been a number of media outlets that ran stories saying this is the moment, this is the time for AI. And as you pointed out, it's not as though artificial intelligence, as a discipline, has been around for hundreds of years. You go back to the forties when Warren McCulloch, Walter Pitts, credited with the birth of AI, they created a computer model of the biological neuron. The next decade in the mid-fifties in 1956, there was that famous conference at Dartmouth College. And as you mentioned, the breakthrough in Transformers just a few years ago. As you look at the development of AI since the forties and there have been media that have said this is it. This is the moment when we will see a transformation, national adoption of AI and it hasn't happened. They're also calling for this moment as being that impactful. What do you think, Roger? Do you think that there's something different now that will see transformation on a grand scale? Do you feel like those tendencies who are saying this is the moment may actually have it right?

Park

I do. I do think so. Just I think with all exponential trends and emerging technologies, everything moves very, very slowly until they move quickly. And that's just the nature of exponential curves, right. I quote Bill Gates all the time. He says, “we always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten years.” And I think that's really informative to the way that we should think about what's happening here, because it's not that there was suddenly a huge spike in AI. AI has been working its way up that curve for the last as you said Jeff, decades. But what's happened is we've finally broken through where I can do things that we haven't been able to do before. So, I actually have posed this question to a roundtable of chief innovation officers. I host a dinner every few months in New York. It's really just an excuse to get folks together and have a nice meal. But we throw out some questions for the group. And I threw this question out. Jeff, I asked the group, you know, what's different about this? And we've been through a number of hype cycles. We've been through a lot of different moments where I was going to be the transformative event. What's different about generative AI? And there were really two good responses to two responses that we kind of coalesced around. The first one was we're definitely at a point to do things with generative AI that we couldn't do before. This is a step change beyond what was done without the technology. So, it's not just about, it's very interesting that an AI can write an email for me and we could do that before. It might have been cumbersome or whatever. The real step changes. We can generate a thousand emails, 10,000, 10,000 emails, 100,000 emails. Each of them can be tailored around a specific theme or tone or targeted at a specific individual. So, the scale of what we're able to do with generative AI now cost effectively, I think that's a massive disruptor and that's one thing that's a change. So, in that way it's very much more like a transition to cloud Jeff, than some of the other technologies that we've been watching, like Web3 and Metaverse and other things. This is very much an inevitable change where we're going to move towards a new platform, a new technology that it won't make sense to once we're there. What makes sense to go back the other way, It won't.

Saviano

And I think you're right, of course, that the technology itself has been rapidly advancing. Of course, this explosion of data, we've been exploring those concepts on the show now for a while. But really for the first time, Roger, don't you feel like that that the superpower of AI is finally in the hands of everyone when and we've already made reference to traditionally, I still have a hard time referring to it as traditional AI. But now we're witnessing the rise of the citizen developer, tools like you mentioned, ChatGPT, democratizing the use of these are these are powerful open source, not just low code, but these are no code tools that are available. Do you think that's a significant factor in this explosion? Is the ubiquity of these tools and how it's now available. It's not that 10% AI team that may have been in the basement of the company. This is really in the hands of everyone, Right? That's a big deal.

Park

I think that's exactly right. Yes. And we've talked about this before. This transition, this inflection point really. From technology as a tool to technology as a platform. And I think we're at that moment with AI and the example I would give is I know folks in the audience might remember the time before the Internet, before the browser, 1996, Netscape Navigator came out. But before the browser, you may remember, there were Internet technologies, there was file transfer protocols. It was a 70% email transfer protocol. There were bulletin boards, you remember bulletin boards Jeff.

Saviano

Yes, I do.

Park

Remember DCP, IP, remember Gopher, remember all these individual technologies that were being used as tools for very specific purposes by scientists, by engineers, by academic institutions or large corporate institutions and it had a specific use for those technologies, and they existed for a while. And then Netscape came out with a navigator, one of the first browsers, one of the 96, and integrated all these technologies together and made it readily available to the public. And really that shift from Internet as a set of tools, Internet as a platform really began. I think the Internet era, the whole digital economy started with that. Another example I would give is the relevance to some of the folks. Younger folks in the audience is mobile phones. I remember like before I had a mobile phone, I had a PDA, a PalmPilot, I had my flip phone, I had a text pager, I had a camera. I remember having cameras and GPS devices and a music player and all these other separate mobile devices that I had on me when I was moving around. And then smartphones came out. I think the iPhone four came out, 2004 smartphones came out, integrated all those mobile technologies together and made them easily accessible to the public. And really that was the inflection point for the whole mobile era. And we've seen the whole mobile economy launch. I think we're in the same inflection point now Jeff, with AI. There are a lot of tools out there that are being used by scientists and engineers for specific purposes, specific applications, like you said, having a team of data scientists and engineers really work on a problem, leveraging the latest AI tools, generative AI and platforms like Judge Unity and Gemini and others are turning that set of technologies from specific tools into a platform that anyone can use. If you can interact with ChatGPT, you can interact with Gemini, you can interact with llama, you get access to the full gamut of AI technologies. And I think we're at that inflection point for what happened with mobile technologies, and we'll continue to do that.

Saviano

I think that comparison is right on. And this is this was kind of fun. I saw this the other day. What's the, here's an AI trivia question for you, Roger. What's the hottest programming language in 2023?

Park

I don't know. Is it a question prompted?

Saviano

Plain English. I thought that was a little smart. And I think it's I think it's so right on, though, isn't it? And you're right. Prompt engineering. Right. But that's what's been so amazing about the democratization of AI is that through prompt engineering, we just brought in one of the leaders in prompt engineering from the team that we aligned to at MIT came in, my friend Daza Greenwood to do some training for us and teach the masses about prompt engineering, and I think we all believe that in another, I don't know, a year, a year and a half, there probably won't be requirements and a need for any, quote, “training” on how to ask these systems. If my favourite part of prompts engineering if you don't know how to ask a question asked the general AI tool, how should I ask this question and it can help you provide an answer. But I thought that was kind of cool that it is. It's all in plain English. And isn't that a driver of how democratized these tools have been?

Park

Right. Absolutely. I'm sorry.

Saviano

You want to keep going? I can just.

Park

Yeah, absolutely not. Yeah. I didn't know you were done. Absolutely. Just I think the democratization of the technology, the access to really, really powerful AI technologies and capabilities and models through a very easy to use interface like chat or visually or through voice, that is going to be a game changer and it's going to be it's going to be very it's already been very disruptive. But I can see the trend line here and with things with the newer technologies that are coming out, the multimodal technologies, it's going to get even more amazing.

Saviano

You mentioned this shift, Roger, from AI tools to platform. Keep going with that. Explain for the audience, if you could, what do you mean by the shift to platforms and maybe give an example of what that looks like?

Park

I think I might have already covered that question with the mobile stuff.

Saviano

Yeah. Is there an example, we can move off of it? Is there something that you could talk about now, a generative AI or maybe even EY specific. I can re-ask the question and be more specific to that. Is that cool?

Park

Yeah. Yeah. Let's do that. Yep. Okay.

Saviano

All right let me, let me do that again. Let's keep going. Roger. With this discussion of platforms and you shared your point of view that we're shifting from a AI tools specific narrow a tools to more of a platform. What do you mean by that? And maybe give an example of a generative AI platform that you're seeing in the market?

Park

Yeah, sure. Tons of good examples. Jeff, I'll use something that we're doing at EY as an example. So, we have quite a few AI tools and capabilities and data management tools, and we curate data, a lot of rules based stuff. We do automation, RPA, all these tools that we have that a lot of other organizations also have RPA automation models, etc. In the past, if we were going to develop a solution or an application that our people could use, we would, you know, we would bring together our engineers, would bring together our ecosystem partners, would collect the data and we do the training runs and we're trying to the models, etc., etc. And we've done that in the past around a very specific scope, business problem or opportunity or issue. But it did take specialists, engineers, it did take data scientists, it did take a very experienced business requirements, folks. And once the application was developed, as you know, there's change management and getting it deployed, training people on how to use it, processes being adapted, etc. We just launched what we call EYQ, which is our generative platform at EI. We have 400,000 employees. It's already been registered by 300,000 of our employees. It's one of the fastest internal tools that have ever been adopted. And through that tool, through a chat interface, anyone at EY can access hundreds of internal skills that we're building and deploying onto the EY.ai platform. And these are they don't have to hire engineers and ask to hire data scientists. They can work through Jeff, as you mentioned, the most powerful programming language in 2023, which is English or another conversational language to ask the question, request the information, navigate our entire catalogue of services and services that are available or get connected to someone else or our ecosystem. All of that through a natural language interface without having to go through procurement to get it, get access to technologies or get access to a team of data scientists. So, that alone is incredibly powerful. The other thing that we're seeing is because its platform based, because a lot of the foundational capabilities are built into the platform, the incremental cost of adding new logic or new capabilities onto EYQ is much, much less than it used to be. You used to take months if not years, to really develop a good AI solution, AI based solution to get the data together and the capabilities. These days, we're rolling out new applications every few weeks, and to me, it's not just the power of the tool that's disruptive, it's how easily accessible it is. And how quickly we can develop and deploy solutions on that platform. That's going to change. That's going to change things for us and for our client.

Saviano

That's great Roger and I, we're all so excited about EYQ here at EY and the Enterprise-wide applications are interesting to me that, that it's the it's the applications that are enabling AI at scale for the masses here. For us close to 400,000 people and we hear this from the clients that we're meeting with, how do you empower the workforce, enable everyone if you're a financial analyst, if you're marketing associate, who needs to draft a marketing communication, if you're a junior attorney in the legal department and your summarizing cases and you're writing legal briefs, or as we were talking about earlier in the innovation process, what if you're product manager, it's that bottoms up value. And I've had a, I don't know, maybe it's a bit of a different point of view, but I feel like the value of generative AI for an enterprise will probably be the highest when you can aggregate the value to those individual roles. For that financial analyst for the marketing associate. As long as these enterprise-wide solutions are enabling at the individual level, isn't that really what it's all about?

Park

That's an excellent that's an excellent point, Jeff. I think we're seeing a lot, certainly at EY and then across our clients, we're seeing a lot of people I would say almost play test with generative AI. They're trying out things or seeing what it can, what it can do, what it can't do. Some interesting applications. We're seeing a lot of things go, quote, “viral”, you know, cool tricks and tips and tricks you can use on the generative AI platforms. What were the real value is going to be to your point Jeff, can we harness generative AI to really help optimize and then beyond optimize, transform our businesses? In order to do that, we need fluency of the technology across all levels of the organization. My view is that I don't think we need to spend a lot of time getting our, you know, younger folks, junior folks, technical folks really excited about the technology. That excitement level's already there, but we can invest in giving them, you know, basic understanding, access to training and materials and access to training and review and coaching capabilities. That's going to be game changing. It's really that middle layer that we have to work through to convince that group of people that you can, AI is here to stay generative AI is disruptive, and you can look at what you're doing today and ask yourself the question, is there a better way that I, with generative AI, with the tools and capabilities available to me now, including generative and but also all the other digital tools and absolutely you can, and finance and risk and legal and supply chain and operations and any type of real business function. I would challenge everyone to look at what you're doing today, what your teams are doing today, and say, can I do this a little bit better using generative AI and then getting access to the enterprise capabilities to build that, deploy that tech, build it, test it, and deploy it at scale. It’s going to be critical. So, the enterprise, the responsibility of the enterprise level is to build those processes of platforms so that everyone has the confidence to really rethink how they're doing their work using generative and with confidence.

Saviano

It's a great way to put it. The confidence is so important, but there's also been an ocean of new and constantly evolving information about AI that's available to business leaders for those in our audience today, Roger and I'd love to have you share your advice on how would you help our listeners discern what's relevant and what isn't, because there's so much information, so many, quote, “experts” telling you what to do and how do you get the value from it? I feel like it's a question that I get all the time. People are overwhelmed if they haven't really dove yet. And to see what's the value to the enterprise and they start to search and look for information, it's almost impossible to cut through it. What advice do you have to people and how do they access the best of what's out there?

Park

I think this goes back to where we started the conversation, Jeff, which is what are the types of capabilities any organization needs to have when they're innovating in a new and emerging space, leveraging new and emerging technologies. And a key part of that is, what are the right ecosystem partners they need to work with? How do you identify them? How do you that the thousands, if not tens of thousands of organizations who are out there touting the latest new technological breakthrough or capability, it just takes time to go through all of them. So, having a good network, an ecosystem of partners that can, one you can rely on, but then you can also leverage to filter all the rest of the information out there is going to be key. The other thing that I think is really important is there, you know, every organization, various levels, certainly we do this, we need to have first party information and insight on the technology. We need to be doing our own testing and learning not to become, you know, experts on everything, although we are building that capability. But I think every organization needs to create labs and sandboxes and play with the technology and really get their own organization hands on with it and creating those places where you can get hands on with the technology, really understand what the art of the possible is. I think that's critical. So, creating those types of experiences or those labs or your own people to test in line with the technology is key. And then also having a forward-looking view on where that technology is going to be. This is another, I think, good, good practice from an innovation perspective, which is we can all figure out what's going to happen today. We can all try to keep up with what's happening today Jeff in AI and generative AI, but it's all going to change six months from now and then since six months after that, it's going to change as well. So, you know, the famous line we got to go to where the puck is going to be, we have to plan out where we think we need to be in the next 12, 18, 24, 36 months so we can make the right investments. Now to be ready for where that technology will be in the future and how and how transformative and disruptive it can be there. So, I think those are all key capabilities that the company companies need to have. And then, I think it's imperative, you know, we say this all the time in the innovation space, but you have to you have to fail fast. The only way to learn something new that where the knowledge doesn't exist or where you don't have access to the knowledge is to try it yourself, test and learn. And if you're going to have to fail 20 times to get to one good idea or one good solution, you should make those. You shouldn't make those mistakes as fast, as cheap and as as safely as possible.

Saviano

I think it's good advice and every company has their own ecosystem and we talk about ecosystems a lot on the show. Certainly, it could be accessing the startup community and there is certainly plenty. I was on a panel last week at NASDAQ with a few venture capitalists talking about AI investing in venture capital, and there is no shortage of money that's flowing into private equity and venture capital for AI solutions. So, and companies will hear from startups all the time. Could be more established. Technology companies, could be a university lab. I love your suggestion about launching labs. You and I both, have both done that and see the value of having a place for those test and learn patterns. And where do you leave it? How do you lead and where do you lead within the organization? And finding that that if you can search through the academic studies, not just a traditional your favourite search engine, but try the scholar approach and we're starting to see now that we’re over a year or so into many of the generative AI tools that interest exploding and we're starting to see some academic studies. There was one from a company called Cyber Haven Labs that actually looked at over a million and a half instances of Chat GPT usage and found many examples of people that were putting company information into these open source tools. So, there is some caution there as well. And that's why having the right governance I think is incredibly important. There are just, there's so many opinion pieces that are at a high level I find that they all sort of read the same. But I'm just curious if you have had luck as well with some of the academic searches and taking it to the scholar approach for searching instead of just a general Internet search for information.

Park

Definitely don't trust everything you see on the Internet. Right Jeff. I think the academics are really jumping. There's a lot of good publications and information coming out of the regulatory bodies as well. As you know, it's a fast-moving space. I think the challenge is there's so much information and having the right ecosystems and partners to rely on to filter that information is going to be key. It really, it is helpful to have a set of trusted advisers in this space to help you navigate all the information that's coming out. And a lot of it. And frankly, Jeff, a lot of this is going to be very competitive If you're in an industry or you're in a sector or function, it does help. I think, to reach out to your peers and counterparts in other organizations to compare notes. Everyone's trying to figure this out at the same time, none of this is going to be a competitive advantage, at least this early on yet. I think everyone is just trying to catch up and keep it together. So, I think the forums like I know you hosted several forums and the forum I hosted on which chief innovation officers, I think those kind of forms are very valuable for people to connect across other organizations and compare notes.

Saviano

I appreciate your point, Roger. How to invest in AI and I think great advice, accessing ecosystems. And you shared a little bit earlier about what does it mean for innovation officers, just as an example from product perspective. But I want to talk for a few minutes about what about within an organization, how AI can be effective in enhancing optimizing existing business processes or functions. Let's use the finance function as, if you have a better example, that's fine. But I thought we could talk about the finance function. How do you feel like generative AI will automate and enhance financial operations as an example?

Park

I think finance is a good use case Jeff, because there are a lot of processes and in finance there are a lot of steps, there are a lot of hand-offs, there are a lot of interactions with other groups and then there's a lot of data and then there's a lot of traceability and requirements around lineage that you have to keep track of and a lot of calculations that require models, but then also some level of judgment or interpretation of that data. So, it's a great, great example. I would say almost every finance process you can look and say is, are there steps in this process that I can automate using AI, are there are there friction points in this process or hand-offs that I can simplify or streamline using generative AI. But I would call that the first level of ambition with AI Jeff. So, a lot of our clients are looking at AI initially as a way to optimize an existing process or system. They're trying to make incremental tweaks, automate steps in the process, reduce some of the friction points and some of the pain points, but absolutely do that continuous process improvement lean, Six Sigma, etc. That's a good application of the technology and an easy one because it fits into a lot of the process improvement structures that already exist in organizations. The next level of ambition and what we're starting to see now just with clients across finance but also across a lot of other functions and industries as not just how do you use AI to incrementally optimize on improving existing process. Now the question is, how do I redesign that process or system to fully leverage the benefits of AI instead of automating steps in a process, can I redesign that process? So that's 100% automatable, and that's where I think things get really interesting because that's not just a ten 20% improvement. This is transforming an entire function, and that's where I think it gets a little bit more complicated in the sense that it's not just understanding the technology well enough to apply it, it's also understanding the processor system well enough so that you can redesign it in a better way using the latest technologies and capabilities. And then one more level of ambition above that Jeff, is once you're once you're beyond optimizing existing processes, once you're you know, you started down the work of redesigning a process around being able to leverage AI better or automation better. The next level is how do you restructure your organization, how do you realign your business anticipating functions that are going to be disrupted or totally transformed in the future? Do you want to get out of certain businesses? Do you want to outsource certain functions? Do you want to be the leader in providing a certain function to it as a utility across an entire industry? That realign ambition, I think, is also going to be key. So not that it's a pyramid, not that you have to go through all these stages, but what we're seeing is that the broad majority of folks that are working in the AI space right now are really looking at that optimize, using AI to optimize existing processes. We're just starting to see some of the more ambitious players in the space start to think about how do you transform our process around AI? And then in some areas where it's clearly obvious that major disruption is going to happen, we're starting to see organizations start to realign.

Saviano

And of course, that has massive implications on the workforce. There are people implications are new jobs that will be created and new roles we're seeing. The most forward-looking companies are the ones that that are, as you said, not just at the bottom layer of that pyramid and looking to automate or enhance an existing process. But how do they redesign? How are they creating new roles? How are they retooling their people to be available and have the skills and capabilities that they need? I get energized, meeting with boards and meeting with senior leaders of companies that feel the obligation to protect their people and help pull them along. And certainly, there are many who are salivating the opportunity to cut 25, 30 plus percent of the workforce as a result of infusing of AI. But I'm really getting energized by the strategic conversations with business leaders who see the opportunity to move their workforce forward in this way. And it sounds like the way that you're describing this evolution of using it internally, that you would agree that massive workforce shift has to happen.

Park

I think there is there are obviously many implications on workforce, workforce planning, the type of work that people are doing and the economics around it. So, just another aspect of how disruptive, generative AI will be in the near term in the midterm, we have I have this futurist that I call hyper automation or hyper productivity, and this was to talk about this before generative AI, but it's even more true or more relevant, I think, in the generative AI space. And the idea is if you can achieve the promise of automation and AI get that 10X productivity, I think there are two things that you can do with that, right? If you're the CEO or the C-level executive in an organization and you manage a big group of people, if you can get ten x productivity delivered from the technology, what would you do Jeff? I think I think there's two things you can do. One of them is you can do everything that you're doing today with 1/10 of people, or you can take all the people and the talent you have now and enable them to do ten times as much. And one of them, the short term strategy and one of them is a long-term strategy. And there's certainly a dip before a leap and a shift in the workforce. And there could be structural changes. But the organizations who can take their best talent and enable them to do ten times as much I think are going to have an advantage. And that and that in hand, Jeff. Another principle that we have, which is called taking the robot out of the human when we have an automation and AI, a lot of what we're talking about is, you know, the question I want to ask is if you've got your people, if you've got your talented workforce doing things that can be automated, number one, it's probably not economically efficient for them to do it if it can be automated. Number two, it's probably not fulfilling knowing that you're doing something that can be automated. But if you can take the robot out of the human and have the automation do all the work, that frees up time for the humans to do human type things. But then I guess the question I want to ask leaders is what are those human type things that are important for your organization? Is it creativity? Is it ingenuity? Is it creating new products? Is it understanding your customers better? Is it working with your other employees better? Is it empathy? You know, is it better collaboration? What are all the things that you if you could free up your people from the monotonous mechanical automatable work, what would you want them spending their time on? Spending more time with customers, engaging them, understanding them better, making your employees more productive and happy? How do you then take the productivity gains that you get from taking the robot out of the human to invest in training and enabling everyone in your workforce to do all the things that they would prefer to do, probably in a more productive way and get that 10x productivity. So, I think there's a there's a very I think I wouldn't call it philosophical, but I think there's a fundamental question to ask first. But I think that question is going to get asked quite a bit in the next few years, Jeff, because what's coming down the pike in terms of what can and can't be automated that equation is changing very quickly with generative AI.

Saviano

I love it. It makes me think of we had a guest on the show last year, David Sconthal. David is a professor of strategy and innovation at the Kellogg School of Management at Northwestern, and he wrote this great book called The Human Elements Overcoming the Resistance that Awaits New Ideas. It's this fascinating view that David has that innovators get too focused on product features and functions. And what you said earlier made me think of this Roger, when today everybody's looking at how do I just add certain features and how can I take out some cost in the system? How can I add generative AI influenced feature to my particular product? What David's point was in this great book is that you also need to highlight and overcome resistance that every great idea has dragged, right? This concept of drag that pulls down from the momentum of a great idea, what it needs to be successful. That comes from a variety of places. David said that it's just inertia sometimes of trying to launch a new idea, could be emotional. He introduced this interesting word called reactance, which is the impulse to resist change. And so, he drew this distinction between friction and fuel. And what you said really made me think of the 10x opportunity is not to just keep adding features and functions onto a product, like that's not going to get you the 10x return. But I was just so captivated by David's approach to you have to focus on the frictions that any new idea has, and we're already starting to see that generative AI is quite applicable to helping leaders overcome those frictions. How do you feel about that? Do you feel like there's an application of David's great work to this issue?

Park

I think I haven't read that book, but now it's on my list because I think it's very workable.

Saviano

That's my holiday gift to you, Roger, that's my holiday gift to you is this book. So please. Yes, that's a good one. You should read that over the break.

Park

Yeah, I will for sure. Yeah. I think this idea that the friction points in an organization and the inertia really right around change management. I know probably everyone in the audience has gone through a large transformation or a large change as experience around how hard it is to get large groups of people to move in the same direction without getting in each other's way. I think generative AI and some of the early use cases we've seen around it and how to streamline the interactions between people are very interesting and they're all, they're all a lot of them are based off of just first of all, just the ease of interacting with and with the generative in it in a very conversational way, a very natural way. The second thing is how quickly that interaction can be tailored based on different styles and different tones and purposes. As you've seen, right? You can go and you can do ChatGPT and you can interact with all these different generating tools and doing what the finance function helped me understand this new accounting role. Help me understand this new process or what's the best. Give me three options for doing X, Y, and Z. But then you can also say, let's speak to me like I'm a five-year-old, or do an acting like you're a cowboy from Western, from the seventies or from the movies in the seventies or talk to me like interact with me. Like I'm interacting with a customer of mine who may be 30 or 30 years old from the Midwest.

Saviano

Yeah, right.

Park

And so that ability to not mimic, but that ability to understand and shift between different styles of interaction and different levels of comfort, to reduce some of the obstacles and barriers to interaction, I think are huge. The other thing that I think comes to the forefront is, as you know, the models are only as good as the data, and the data is only as good as how well you can curate that data. The importance of curating data so that it's accurate, complete and unbiased and represents the tone of your business, I think is interesting. And what you referenced, Jeff, one of the things that we're talking with clients about is and one of the one of the natural applications of generative AI is contact centre interactions. How do you streamline the contact centre interactions? How do you make sure it's compliant? How do you generate responses that are going to be more effective, etc., etc., here and there? Here's a next level question. Once everyone rolls out their generative contact strategy, Jeff Here's a next level of on top of that, which is can you make that contact centre agent feel like that as an employee of your organization to your customer?
Right. So if someone interacts with an EY chatbot, we're not supposed to call them box anymore, but someone interacts with an e y generator, the agent can we, can we figure out a way for that person to feel like, I know that was an agent, but I can tell it wasn't an agent because in a tone, because of that the pace, because of the level of respect or understanding or the level of talk, those are, I think, the next level of what we're going to see with generative AI and if we can get those agents in place, those systems in place, that level of understanding in place what and how people can afford to be, how people prefer to react and how they prefer to be treated. That I think, will reduce a lot of the cultural friction when it comes to transformation, for sure.

Saviano

I think it's that, I think you hit it on the head. Roger. The opportunity to hyper personalize your solutions. And we're seeing this in many different sectors, the sector that you were closest to for years, financial services. Opportunities to offer a range of services that are actually targeted to specific customers. So, your customer feels like this is personal to me and the company really understand us, understands the needs that we have raises a lot of ethical issues about. Do you feel an obligation to tell your customer when the hand of eye has been used in a particular product? And we're certainly seeing companies addressing these ethical issues many times. These are first impression and AI ethics issue can be quite thorny. Ethics is about asking questions, and it's about resolving tension sometimes between two competing priorities. So we're going deep and deep into air ethics because frankly, there's not a lot of legal requirements yet in the world that companies and other on the development side, but also on the application side as those laws and regulations develop. Our point of view is that ethics is where it's at today, but that companies need to pay attention to the ethical issues around AI. Roger, we could talk, you and I could talk about AI forever and ever. We have a special feature on Better Innovation. We have three questions that we close out every episode. What do you think?
You up for it?

Park

Yeah, let's do it.

Saviano

All right, here we go. First question, What's a book that has greatly impacted you? So many books. So many books. And as you know, I'm an avid reader. I would say in the context of this conversation, two books come to mind right away. One of them is Superintelligence by Nick Bostrom, which is pre the generative AI way, but very much talked about the possibilities and dangers and hazards of artificial general intelligence. So great book. Nick is a philosopher, not an AI scientist, but a great perspective. And the second is a book called Finite and Infinite Games by James Carr's, and it's about mathematical ideas around game theory, finite games versus infinite games. And I won't spoil the book for you Jeff, but the one thing that stands out to me is a finite games you play to win an infinite games you play to continue the game. So, understanding what type of game you're playing really influences your strategy in this space.

Saviano

And I'm so glad you mentioned superintelligence and I have it. I haven't read it yet and have had a few people recommended that this discussion of the hazards of AI is not just for the few who are developing large language models, but in our travels with companies and meeting with boards that you know, we've been making it clear that that even companies that are applying this powerful technology that there are ethical and other hazardous as the word is important, hazardous considerations great, great advice.
We're going to hopefully we'll get this episode out just before the holidays and people can buy these books for their favourite family or friend they’ve got to give a gift to. Okay, here we go. Next question. What piece of advice would you give to a younger version of yourself?

Park

Probably take better care of my knees.

Saviano

That I guess the gentleman that I was I was expecting the deep, deep Roger Park answer and I got knees. Okay, that's I wish I had taken better care of my knees.

Park

The best professional advice I'd give myself is probably to choose the jobs and projects where you're going to learn the most rather than make the most. I would say looking back over my career, just the real turning points for me have always been where I took a job or project where I was a real stretch for me versus the most glamorous project in time.

Saviano

Great advice, especially those early years. Yeah, don't worry about how much you're making right off the bat. Just learn.
Have great mentors around you. Right? And great advice. Okay, last question, Roger. You're doing great. Here we go. What areas or industries do you feel like are ripe for innovation in the next 3 to 5 years? This one was especially tailored for you, Roger. What's ripe for innovation?

Park

Well, I would say financial services for sure, because that's the industry I'm closest to. In general, though Jeff, I would say any industry that's in the business of creating content or creating the environment that generates content, because in a lot of ways content is going to be the new capital in our digital economy. And then to the point you made earlier, any industry that's at the intersection of hyper personalization and trust is going to be ripe for innovation over the next few years. And I would put financial services in that category in a lot of other industries for sure.

Saviano

Yeah, we're probably spending two thirds of our time in financial services because they more so than other sectors. They understood traditional AI. They've been using AI for years. It's not their first AI rodeo. And so, it's been an easier shift. Roger, I really appreciate you coming on the show.
You are always welcome anytime you've got another fresh set of ideas which is probably be like next Tuesday afternoon. You're always welcome on the show. So, let's continue to stay close and I really appreciate you spending some time with our audience today. Thanks so much.

Park

Jeff. It's been my pleasure. And I will come back on as often as you'll have me.
It's been it's been fantastic. Thanks, Jeff.

Saviano

Terrific.

Presenters

Jeffrey Saviano

EY Global Emerging Technology Strategy & Governance Leader

Podcast

Season 7, Episode 8

Duration

57m 36s

In this series

EY refers to the global organization, and may refer to one or more, of the member firms of Ernst & Young Global Limited, each of which is a separate legal entity. Ernst & Young Global Limited, a UK company limited by guarantee, does not provide services to clients.