EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
This episode of the Think Ecosystem podcast features host Sameer Shah, EY SAP Technology Partner, in discussion with Dan Diasio, EY Global Artificial Intelligence Consulting Leader, and Nicolas Babin, an expert in artificial intelligence (AI) and digital transformation. Together, our panel delves into the potential AI bubble – the disconnect between inflated expectations and actual capabilities – and what a pop could mean for organizations.
Listen in to learn more about AI’s evolution and to hear actionable insights on how companies can prepare their data to take full advantage of AI’s potential.
Key takeaways:
Concerns about an AI bubble reflect market sentiments, but the real opportunity lies in leveraging AI's capabilities thoughtfully. Even if progress slows, AI remains a powerful tool that organizations have yet to fully exploit.
AI should augment human capabilities, not replace them. A collaborative approach that integrates AI into workflows enhances productivity, builds confidence among employees and fosters a culture where AI empowers the workforce.
Executive involvement is critical in driving AI initiatives. Leaders need to foster an environment that encourages experimentation, embraces change, and aligns AI strategies with business objectives.
For your convenience, a full text transcript of this podcast is also available.
Announcer
Welcome to the EY Think Ecosystem podcast, a series exploring the intersection of technology, collaboration and innovation. In each episode, we orchestrate insights and stories and perspectives from across the EY Partner Ecosystem, our own client base, our leadership team to address the most important issues and challenges of today.
Sameer Shah
I'm your host, Sameer Shah, EY SAP Technology Partner based in the UK, specializing in helping to shape and deliver business transformation through the use of modern technologies. Artificial Intelligence stands at the forefront of technological innovation in today's rapidly evolving business landscape. From enhancing customer experiences to optimizing operational efficiencies, AI promises to revolutionize the way organizations function. Yet amidst this excitement, there's growing chatter about an AI bubble, a potential disconnect between inflated expectations and actual capabilities. So, what does this mean for companies if they hear a pop? Is it a signal of diminishing potential or just an opportunity to recalibrate and realign strategy for growth? More importantly, how can organizations prepare their data, the lifeblood of any AI initiative, to harness AI's transformative power? In this episode, we'll delve deep into these questions, and we'll explore the historical context of AI's evolution, demystify its hype and provide actionable Insights on how businesses can prepare their data for the AI era.
Shah
We'll discuss the critical role of strategic alignment, the importance of robust data governance and the human-centric approach necessary for successful AI adoption. Before we begin, please remember, conversations during EY podcasts should not be relied upon as accounting, tax, legal, investment or other professional advice, and listeners must consult their own advisors. I'm very excited to introduce our guests today. Dan Diasio, EY Global Artificial Intelligence Consulting Leader, is joining us. Dan supports clients with AI-enabled business transformation, focusing on the strategic direction, design and deployment of responsible AI and modern data platforms.
Dan Diasio
Thanks, Sameer. Very happy to be here with you.
Shah
We also have Nicolas Babin, a renowned expert in artificial intelligence and digital transformation with over 25 years of experience. Nicolas launched one of the first AI-based robots in 1999, and he also serves as a digital ambassador to the European Commission on New Technologies.
Nicolas Babin
Thanks, Sameer. It's great to be here. Thank you.
Shah
Thanks to both of you. Let's begin by understanding the current AI landscape. If I come to you first, Nicolas, AI has been the subject of fascination and innovation for decades. Given your extensive experience in the '90s, how have you seen AI evolve and what key lessons can businesses learn from its history?
Babin
Yes, great question. Thank you very much. Actually, I've seen AI progress significantly from what we call early-based systems to today's advanced machine learning models. My experience when we launched aibo back in 1999, as you said, we showed that success depended on high-quality data and a very clear understanding of AI's role in business. But the evolution of AI has been marked by several waves, each pushing the boundaries of what technology could achieve, from rule-based systems in the '90s to the sophisticated machine learning and neural networks that we see today. Companies today should see AI as a long-term investment, focusing on data quality, strategic alignment and ethical use. This approach will enable sustainable, impactful AI adoption that will align with business goals, and that's what we are all looking for.
Shah
Thank you, Nicolas. Dan, building on that and your work across clients in various industries, how do you perceive the so-called AI bubble? Do you see it as a cause for concern or an opportunity for businesses to start to think about how to leverage AI more effectively?
Diasio
Yes, Sameer. Thanks for the question. I go back to where Nicolas was describing these different waves. We're in this new wave that is really probably largely shaped by the invention of the transformer and the repercussions and the scaling of data around the large language model. I think there is a lot of talk about a bubble, and there's two things can be true at the same time. On the one hand, many organizations that are in this AI arms race are investing billions of dollars in collecting and processing data, spending a lot of money with cloud computing and standing up new and modern data centers. Aside from ChatGPT, many companies are still looking for the best product market fit that justify all that investment. That's really the talk about the bubble. There's an incredible amount of investment in this space, and AI hasn't yet started to deliver a return against that level of investment. That's just on the one hand. On the other hand, this is a very powerful capability, even if the pace of innovation doesn't continue up the S-curve that it's been over the course of the last 36 months. Really, most of this technology is not being leveraged to its fullest extent.
Diasio
When it comes to the AI bubble, I would say even if progress starts to plateau, this is still largely untapped within most organizations. The future growth of AI should not stop any organization from figuring out how they can look inside their business and leverage this more effectively than what they may be doing today.
Shah
That's some great insight, Dan. Building on that, Nicolas, considering some of the current hype around AI, what are the common misconceptions you've been observing in the industry and how should organizations address that?
Babin
I've seen so many misconceptions since the late '90s. It's interesting. But many people think that AI can solve business problems instantly or replace entire roles. But AI is best seen as a tool for augmentation, and this is really a point that I always made about human being augmented. Rather than viewing AI as an end all, organizations should aim for clear, achievable objectives and use AI to enhance human strengths. It's a bit what Dan just mentioned in the fact that it's untapped yet, and that's one of the reasons why. With a balanced collaborative approach that will build trust and maximize productivity, we'll be able to focus on ethical practices and we'll ensure that AI initiatives are sustainable and beneficial to all stakeholders. That's a really important point here. Also, if I could say with the SAP/BTP vision, we have the future of AI that lies in human-augmented systems where technology enhances human capabilities rather than replacing them. So to my first point. Probably if we agree to call it a AI bubble, but the burst of the AI bubble may actually serve as a wake-up call to all refocusing efforts on meaningful sustainable applications and going to the untapped world and tap it into it.
Babin
If I can, bounce back on what Dan said.
Shah
Thank you, Nicolas, for that. Dan, just one more question for you in this part. What trends are you seeing within the strong EY Ecosystem and how are our clients exploring AI and across business processes?
Diasio
Yeah, I think many organizations, if you look at how generative AI was introduced to the public and to many organizations, it was in the form of a chatbot, where you ask questions and you've got answers. While that's an incredibly powerful capability, many organizations are finding that that capability is not alone what's going to really drive this transformative force that AI is. They're shifting from a mindset of looking at answers – looking at a system that can provide answers – to a system that can provide actions. That is really where you start to hear this movement of agentic AI or AI agents. These are now using the capabilities to actually do some of the work and persist over time, as opposed to just providing an immediate response to an answer. That is really a result of many organizations finding incredible potential but not being able to get the return on the chatbot, question, answer type of construct. What we see many organizations doing are looking at their portfolio of where they think AI is going to drive value. Instead of looking and figuring out how you'd prioritize 300 or 400 use cases, they are now starting to think, How can I embed these capabilities into the workflow?
Diasio
That might mean a more transformative mindset, thinking about these as reworking and reimagining a business process. There, reasonably, it's no longer a list of a couple of hundred things, but it's a handful of initiatives. We see this big move of going from a lot of small use cases to a couple of initiatives. Those initiatives really start with reimagining the flow of work and inserting AI into that flow of work.
Shah
Thanks, Dan, because that leads us really nicely onto the topic of data and strategy. Data is often cited as the foundation for successful AI initiatives, and you've mentioned process there. What challenges do organizations face in preparing their data for AI? What best practices would you recommend for overcoming these obstacles?
Babin
To me, data preparation has evolved significantly from what we saw in the late '90s, data cleansing and structuring to more complex today purpose-driven processes that incorporate business rules and cater to agentic AI, as Dan mentioned. In earlier projects, data preparation was definitely primarily about ensuring accuracy and eliminating redundancies. That was really what we were focusing on. Whereas today, what I see with my clients is that requires an understanding of the business logic and specific context in which AI will operate. For example, in customer service, we talked about chatbot, and Dan mentioned it, and it's very true about the use of chatbot. In customer service, they need to tag and structure data with precise business rules, so responses align with organizational policies and customer expectations. Additionally, we see that unstructured data, such as emails, documents, images, you name it, is now increasingly valuable. Organizations should invest in techniques like natural language, processing and image, labelling, to extract meaning and structure from this data. That's one of the points that to me is essential as AI systems learn and adapt in real-world environments.
Diasio
I couldn't agree more, Nicolas. I think about it, if we were to go hire an expert, an industry expert that was one of the best and most renowned in their field, and hire that person into our organization, you probably don't have that individual making decisions on day one. First, you probably give them an overview of what the firm's strategy has been, how the firm is organized, how the company works, what's some of the ways of working, the investment profile, et cetera, before you ask them to make decisions on behalf of the company. This is what many organizations are finding with AI. They're using a public model that are available and trained on the internet, but before they start using them to make decisions or to help shape recommendations inside the company, you need to give it organizational construct. There's a lot of technology terms that people use, retrieval, augmented generation, using knowledge ontologies. But these are ways of being able to capture what you would want to teach somebody about the company before they start making decisions. The interesting thing, and Nicolas, this is picking up right on your point, is that most of the intelligence or most of the context that you would give that person is not sitting just inside the data lake, but it's a lot of the knowledge that exists across the organization, the culture, the way the company makes decisions, some of the tacit understanding and policies and guidelines and standard operating procedures.
Diasio
For most organizations, that is not a focus of their data strategy. The data strategy is mostly focused on the data lake and master data management. For many organizations, I'm finding that we need to actually expand the definition of data. It's no longer just data, but it's data and knowledge. That knowledge strategy is particularly important as it relates to this generation of AI.
Babin
Couldn't agree so more with you. It's about governance policies. That's exactly what you're describing here. That's the key point in order to move forward with this.
Diasio
One of the things I often find is that when a company builds, let's say, a RAG system, and they give a lot of context about how they run their process off to a storage mechanism, a vector database, or what have you, they often dump it into the environment once, and then they have the same problem that they've had with master data for a couple of decades. With now, you need to have somebody that actually manages the knowledge and expertise, and many times that's a new skill. I think we're on the beginning of a new frame of what data management means for organizations, and a lot of those best practices are being defined today. But there's a lot that they can get from looking back in history.
Shah
That's great, both of you. I think we're fully aligned on the need for a definition of a data model now that makes sense, a wider definition, perhaps, than before the need for data preparation, the need for data quality and readiness. Speaking of preparing for AI in the same vein, Nicolas, we've talked about data, but human-centric AI, how can you elaborate further on how businesses should ensure that their AI initiatives are designed to augment human capabilities rather than replace them?
Babin
That's a subject that's very close to my heart. I have to tell you a story. Back in 1996, we're in Manchester, the Trafford Centre, one of the largest mall in Europe, I think, or probably in the world. It's huge. I had taken aibo, the first AI-based robot. It was a prototype, and I took it out to show people how it worked. One guy jumped on it and ripped his head off saying, I don't want robots to replace me. That really hit me at the time thinking, I need to be very careful. I'm starting a business here where people are worried about this. So human-centric AI is about using AI to enhance, not replace human capabilities. That has been in my pitch, my words that I've been using since 1996. Businesses can achieve this by integrating AI into workflows to handle repetitive tasks while human focus on complex problem solving. That's really one of the key. With transparency and interpretability we’ll help employees have more confidence in AI and use it effectively. With this collaborative approach, we will be able to foster culture where AI is valuable tool that empowers employees. Two very key words here is the culture, because this is also about changing the culture, but also empowering employees.
Shah
Thanks, Nicolas. It sounds like we have the data question, preparing it in the right way, and perhaps looking at the definition, we've talked about the need for ensuring this is human-centric approach. Dan, in terms of the EY-SAP Alliance, that's quite a uniquely positioned in the market to work with organizations to provide long-term value. With all these aspects that we've just mentioned, what's the right balance between innovation, regulation and meaningful business outcomes?
Diasio
Yeah, it's a really big question, Sameer. I'll take my best shot at giving you a punchier answer. But if you think about what SAP is, it's really at the center of most enterprises and really the center of process and workflow. It's uniquely positioned to be able to start to leverage a lot of the business rules and the security parameters and the workflows to be able to really power some of these processes in the future. One of the things I find really beneficial about the EY-SAP Alliance is that together we are focused on not just using AI to do work or to do migrations more efficiently, but also working together to figure out how could these processes in the future best be powered by AI? And what is a meaningful and long-term data strategy that helps companies understand and make sense in a future resilient way where data should sit and how it can be best leveraged for AI systems as they continue to turn on in the future? So I think there's a lot of collaboration, not just between our sales teams, but between our product teams in shaping where and how things can evolve in the future and what our clients need, not just today and tomorrow, but a year from now or two years from now.
Shah
Nicolas, just to add to Dan's thinking there, in the context of change management, what strategies should organizations employ for successful adoption of AI, especially considering that cultural shift and that employee engagement experience that's required?
Babin
Yes, again, another great question because as we just mentioned about the cultural change, it's really important. A adoption, to me, and what I see with my clients, is that it relies on clear communication, helping employees understand benefits of AI and purpose within the organization. Early engagement, such as involving employees in pilot projects, promoting a sense of ownership and building enthusiasm is really something that I see are key. If you continue training, ensuring that teams feel confident and prepared to work alongside AI, this is another element that I've seen it being very successful. Creating a culture of adaptability will make AI integration much smoother and more sustainable.
Diasio
I think we need to focus on potential and not just productivity with AI. Because as you're saying, Nicolas, this is about empowering people to be able to do more. If the narrative is just, how can you use AI to be able to do this work more efficiently, and we're not completing the sentence of, with that efficiency saved, now, how can you either use AI or redeploy your time to do work that had not been possible before. When we do that, we are now empowering people. If we fail to do that, we are just in a way automating their work. I think that's really important to the change management story that we're asking people to complete the sentence and we're empowering them to go do more. Otherwise, we're going to be stuck in a world where there's just too much overly focus on productivity and not enough on what the new potential can be. But this is all nice from a concept perspective. Sameer, I want to turn it back to you if I can play the host for a moment. Do you have any examples where you're seeing this with respect to your clients or your work in the field?
Shah
Yeah, absolutely. I think what you guys have been touching on is absolutely key. We touched on data, we touched on process and we touched on the people agenda and ensuring adoption. For me, these functions are intrinsically linked, especially the data and the process flow perspective. Because if you define the end-to-end process and you define the interlinkages based on that, and you've got an enterprise data model, you have the best perspective for understanding the role of AI as a contributor to value. Let's talk about a real-life example, Dan. If you take the usage case of asset health monitoring, i.e. the analysis of historical and current data to predict and go forward looking on product defects. That is in turn linked to supply chain planning, maintenance scheduling. In turn, that's affected by employee working hours and skills. There you have a scenario that's really across asset management, finance, supply chain and HR. To get the value you need, you have to have the right data at the right time and a high level of data quality behind that. Especially as organizations look more towards agentic AI, that for me is how you gain the value through cross-collaboration.
Shah
For me, it's absolutely critical. It's also critical from a change management perspective. When we've run events in the past on the potential for AI, having multiple businesses and functions and process owners in the same room has really been the key to unlocking value because you then gain a common and clear understanding of the potential for AI across a group of people. You have better alignment on the need to define a strategy for AI, which everyone is bought into. Then you also have higher potential for agreeing on the priorities and the sequences and the funding. This concept of cross-collaboration, for me, we need to move away from silo-based AI POCs. I mean, they have a place, they absolutely have a place. But cross-collaboration across interlinked processes and an enterprise data model, that's where we're really going to tap into value. Just coming back onto the people agenda, Dan? In terms of leadership, in your experience, what role does leadership need to play in driving AI initiatives? And how can leaders foster an environment that's conducive to maximizing the potential for AI?
Diasio
Yes, Sameer. I Love this question. This form of generative AI is no longer just about data scientists. This is a complete participation sport by every member inside the organization. My guess is that at most organizations, the C-suite is the most likely to have been presented with all the opportunities and potential that AI can use to help power their business, and are probably the least likely to actually have time in their day to experiment and use it. Leadership is extremely important, and it's so that they can see how it works and experience it as opposed to just know the narrative. One of the things we've been doing to address that with clients is to run executive training sessions with our clients. We call them our AI master classes, and we've done this with many organizations now. I just want to describe a little bit of what that looks like. We had a session about two weeks ago that I was leading with a CEO of an industrial company and his executive team, and he brought them all together. We had two and a half hours. We could have done the typical long-listed presentation of slides and just go into different sections.
Diasio
But in this master class concept, we actually give them an exercise and we bring laptops where they work in teams and put their hands on the keyboard and do the work. In those moments, it changes. It goes from a vague concept to something that they now can see and figure out how to best apply in their business. In that particular instance, the company commissioned one of their leaders to go drive a strategy for AI and then actually said, we are going to focus with our customers as well, and we're going to run similar joint sessions with our customers so we can find ways to create mutually beneficial and joint value. But that leadership tone at the top is critical because it's a lot less of a system implementation and a lot more of the beginning of some sort of a transformation. Unless if you have the permission from the executives at the top to think about doing work differently and to think about engaging with your employees in a different way, as opposed to just implementing some technology, these things are very difficult to get off the ground. Leadership is one of the top or maybe one of the top two characteristics of companies that are off and running with respect to AI in their business.
Diasio
Sameer, with your work with clients to achieve transformative success, how can businesses leverage AI capabilities to drive more intelligent decision making across their enterprise? What's been your experience in making that real with your clients?
Shah
That's a great question, Dan. I'm really lucky because in this respect, my work mainly centers around business transformation enabled by SAP and surrounding hybrid architectures. At the moment, there are some huge programs being defined as our clients move to a digital core. These programs are typically used to optimize business processes, reengineer data models, re-architect analytics strategies, etc. They really present the ideal opportunity to look at and leverage AI as well. What I'm seeing, Dan, in the market is clients are really benefiting in several ways. Firstly, they can start to leverage the AI capability that vendors are building natively into their products. So for SAP, for example, that could be Joule and/or Embedded Capabilities in finance, supply chain, et cetera. But they can also benefit from solutions which SAP partners, such as EY, are developing. And on top of that, you have SAP BTP, which Nicolas mentioned, the business technology platform, which is a really smart thing that SAP have done here. They've made that an open hub. So that plays very nicely into the integration into a wider ecosystem through SAP's AI partnerships. So there's just a number of ways, Dan, that I'm seeing in the market.
Shah
You have the transformation programs that are looking at data and processes, and you have native and new capabilities being built up by ourselves, by our clients, by our Software Alliance partners. There's a really great momentum in the market for doing this at the moment, and the timing seems very good to take advantage of this. Let's move on to real-world applications as this is moving very, very quickly. Could you both share a few examples of organizations that have successfully prepared their data for AI and the tangible benefits they've seen as a result?
Babin
Sure. If you don't mind, I'll start then. I've been involved with a very large healthcare project in Dubai. That's the only thing I can say about the project, where a major healthcare provider focused basically on improving diagnostic accuracy by implementing AI to analyze patient records and medical histories for early disease detection. This has been one of the major projects that I've seen, and I was really impressed with this because to achieve these, they invested heavily in data preparation, which goes perfectly well with what we've been discussing from the beginning. They standardized patient records, they cleaned data to remove inconsistencies. They ensured compliance with strict healthcare regulations. As you know, it's one market that's very, very regulated. With this structured approach, it enabled them to deploy an AI model that could reliably flag early signs of diseases like cancer or heart conditions. The result that we saw was noticeable reduction in diagnostic errors, which is one of the key issue in healthcare today, leading to earlier intervention for patients. As you know, the earlier you can intervene, then the better for the patient. Then, consequently, we improved treatment outcomes and patient satisfaction. It was really impressive project, and it was just last year.
Babin
It's something that's very recent.
Diasio
Nicolas, I love that example because I think that is really addressing potential as opposed to just productivity, as we were describing earlier. We're working with a number of consumer product companies that are rethinking about how they do their R&D and how they come up with product design. Instead of traditionally going through this process of collecting insights from an R&D perspective and then coming up with designs, they are now supercharging that with AI so they can get a lot more ideas through the product innovation and design process. We've even seen some companies deciding to do dynamic focus groups, leveraging different personas in generative AI to be able to start to get some real-time feedback. They're removing and minimizing many of the ideas that are not going to be addressing consumer needs in the short term and are able to disproportionately focus on the long term. Some of their early results are that when they go through this process, they are 10 to 15% more revenue-generating than those that haven't gone through that process. There's also a number of companies that are addressing and empowering employees in jobs that tend to have quite a bit of turnover. I think many people know that working in a call center is very difficult work, and it can be very stressful, and often those jobs have pretty high rates of turnover.
Diasio
What we see with one of my clients is that they wanted to not replace the worker, but instead close the time that it takes for that worker to be able to take on a wide variety of different call types and cases. Instead of going through a nine-month training program, now they can have an AI assistant listening in on the call and helping direct them to work through complicated tasks so they can do more and resolve more on their own without having to transfer to other people. That's been able to both really drive job satisfaction as well as significantly reduce call time, significantly improve what the client's experience have been. There's this broad continuum of examples of where companies are using it. I'd say, just to a point that I made earlier, that there's no shortage of ideas. Most companies at this point, we have a statistic that we ran as part of our AI Pulse survey that says 96% of companies are investing in AI at this point. Many of those companies actually have something in production. But what we're finding is it's those ways of rethinking the work which are really driving the real tangible benefits.
Diasio
When you do that, you also have to face off against some of the challenges from a data perspective as well.
Shah
That's great, Nicolas and Dan, some real value-led scenarios there. I really enjoyed hearing about those. As quickly as things are moving now, looking ahead, what emerging trends are you seeing in AI that organizations should start thinking about, start preparing for, the ones that will have the most transformative impact on businesses in the next 3-5 years.
Babin
It's always hard with AI and new technologies to think about 3-5 years because it's a very long time away. But let me have a shot at this. One key trend which goes into what Dan and I've been talking about AI agents, agentic platforms. It's what we see the growing sophistication of generative AI. Beyond just text, generative models are advancing in areas like image, audio, video generation. I've seen that as well, opening up new possibilities for personalized marketing, content creation, and even product design. I've seen companies start building structured and quality data sets now, as we discussed, and that will help with generative AI's output because generative AI output is only as good as the data it learns from, obviously. I tell company that they should also invest in talent and partnerships to navigate this technology responsibly, ensuring the generative outputs align with brand integrity and ethical standards. That's one. The second one that I see more and more and that I think will be very successful in the next 3-5 years, or at least mainstream in the next 3-5 years, is AI-augmented decision-making. AI will increasingly serve as a partner in strategic decision making, particularly with advancement in explainable and interpretable AI, which will help clarify how models arrive at conclusions.
Babin
So organization will be able to prepare by refining their data governance frameworks, ensuring that decision making data is accurate, unbiased and reliable. Just want to say, at least with my hat of ambassador to the EU, here a digital EU ambassador, we work very hard on this, making sure that the data will be accurate, unbiased and reliable. We've created a CE label that will help companies to prove that that's the case in terms of the quality of the data. The third one that I would see, the third point that I see is Edge AI. I see a lot of Edge AI being poised to reshape industries by processing data closer to where it is generated. We see Edge Cloud with Edge everywhere. But here, it will enable faster real-time insights without reliance on central cloud infrastructure. Industries like retail, healthcare, manufacturing, this can lead to faster, more responsive operations. And organization will evaluate how they can integrate Edge AI within their IoT devices and infrastructure, and then they'll be able to focus on optimizing their data architecture for decentralized processing. The last point in terms of the trends that I see is ethical AI. It's a trend, it's not a product, but here it's a trend that more and more, and I'm really glad to see that because ethical AI and responsible AI practices are trends that are rapidly moving from optional to essential.
Babin
I see that every day. Dan, I'm sure you'll have a chance to talk about that, but it's really something that we see regularly. We have regulators and consumer alike that are increasingly concerned with transparency, fairness, and data privacy in AI. Organizations that establish robust ethical frameworks from data collection to AI will not only comply with emerging regulations, but it will also build trust and goodwill with consumers and stakeholders.
Diasio
No, I agree. I just want to add on to your last point there, Nicolas, because I find that when you have to continuously build confidence in AI, what we are seeing is the emergence of a trend of moving away from the PowerPoints, the frameworks, the governance processes, and actually starting to build tools that can help build better guardrails and build better detection of hallucinations or when there are risks of a prompt injection attack. I see a lot of the responsible AI and confidence in AI initiatives going from PowerPoint now to code. In that essence, that starts to take those capabilities and start to push them down where the control points are happening. But I fully agree on your point, some call it agentic AI, but I do think it's going to very much evolve into much more real-time decision-making.
Shah
Thank you both. Some amazing insight into the trends there. If there is one piece of advice that you'd give to organizations who are hesitant about investing in AI due to the fear of the bubble bursting, what would it be and how can they move forward confidently?
Diasio
Yeah, Nicolas, maybe if I take this one first. Let's imagine that the AI gets no better than it is today. It's still the extremely powerful capability that has PhD-level recall, and notice I didn't say intelligence, but PhD-level recall that you can spin up for a dollar. We've not figured out how to use that. If the bubble bursts and explodes in some a colossal fashion, that is still a tool that we have not figured out how to use. I'd say for organizations, it doesn't matter in a way what happens with the state of the art because it's very difficult, as Nicolas said, to predict 18 months out, let alone three years out or five years out now. But for most organizations, if you're not rethinking how you deliver value to your consumers or your customers or your stakeholders or what value it is that you're in the business of providing, if you're not looking at that very essence, then you're falling behind. I wouldn't be distracted by the bubble and the financial pressures that are happening. I wouldn't be so worried about what's happening in the finance space with respect to stock prices and investments.
Diasio
Organizations need to start looking at how this can really impact their business models or their operating models. That is still applicable, even if the AI doesn't get any better than it is today.
Babin
Very, very good point. If I could add – to me, organizations that are concerned about an AI bubble should not focus on the concern, frankly. I totally agree with what Dan said, but you need to focus on AI as a long-term strategic asset rather than a short term trend. Ai has not been a trend. It started back in 1950 when Alan Turing starting to do his Turing test. It's something that's been around for over 75 years. AI's evolution over decades has shown its resilience and adaptability, because we're still talking about it and we see it now every day, everywhere in our lives, making it a transformative force when it's applied thoughtfully, and that's a very important word. To move forward confidently, they'll need to start with targeted, manageable AI projects that address specific business needs, allowing them to see measurable benefits early on. When they approach AI incrementally and investing in data quality and ethical practices, an organization will build sustainable AI capabilities and deliver genuine value, regardless of market hype.
Shah
Thank you so much, Nicolas and Dan, for sharing your valuable insights and experiences. It's evident that while the AI bubble might reflect market sentiments, the real opportunity lies in strategically preparing our data and aligning AI initiatives with business objectives. For those organizations willing to invest in robust data infrastructures and embrace a human-centric approach, AI offers unparalleled potential to drive innovation and growth.
Babin
Thank you very much, Sameer, and it was great to participate in this.
Diasio
Thank you, Sameer and Nicolas. This was fun.
Shah
Before we go, I will just say that the views of third parties set out in this podcast are not necessarily the views of the global EY organization nor its member firms. Moreover, they should be seen in the context of the time in which they were made. I'm Sameer Shah. I hope you've enjoyed the show and that you'll be able to join us again very soon for the next edition of the EY Think Ecosystem podcast.
