Webit 2026: AI strategy is no longer an innovation topic - Dessi Sevdina

In this talk, Desessie Sdana, managing director at Marsh, discusses the integration of artificial intelligence into business strategy, emphasizing its transition from an innovation initiative to a core aspect of corporate governance. The speaker explains that with AI becoming a board-level priority, organizations must adapt quickly and align their AI strategies with operational goals. Sdana shares insights into the complexities of implementing AI in large corporations, including dealing with EU regulations and the importance of workforce transformation. She highlights the need for organizations to balance internal and client demands while ensuring compliance and risk management in their AI applications. The talk also addresses how AI can enhance productivity without eliminating jobs, focusing on optimizing workforce tasks.

After looking at how AI is reshaping teams and workflows, we now move to a broader leadership question. What happens when AI [music] stops being an innovation initiative and becomes part of business strategy? Please welcome Desessie Sdana, managing director and head of strategy at Marsh with AI strategy is [music] no longer an innovation topic. Desi will explore how AI is becoming a board level priority shaping leadership, governance,

workforce transformation, risk management, resilience, and operational continuity. Because in the AI era, the leading organizations will not only be those with the most advanced models. They will be the ones that adapt fastest, align best, and turn AI strategy into real execution. Please give a warm Webbit welcome to Desi Svdina. Hey. Uh, good afternoon everybody. I'm going to do my presentation in English. Although I am fully Bulgarian.

I haven't lived in Bulgaria in probably over 20 years. So, I'm very, very happy to have been invited here. and talk about the future uh of this country, technology and where AI fits into it. So, um you may see a slightly controversial title that AI is no longer an innovation topic. That's because we already all use it, right? I think we've heard so many companies today from

startups to medium to very large global enterprises already have embedded it. So for me as head of strategy for the European region, I just don't see this as an innovation topic. It's a topic of how you're going to differentiate yourself from your competitors or how you're going to deliver value to your clients. So some of you may be designing AI, some of you may be designing software

that you're trying to sell to a very large corporation like mine. So, we're a Fortune 500 company based in the United States, but we operate in 130 countries. You probably don't know our name, but we probably touched every industry out there. We're the largest risk advisor in the world. Um, we offer management consulting, risk consulting, human capital consulting, and investment consulting. So, we typically advise 16 trillion

dollars worth of assets. So, for us, um, When we buy AI, it may feel frustrating to some of you who are trying to sell it to us. But I'll go today through why that is. And some of you may work for really, really large corporations like mine, and you may be frustrated like sometimes I am as to why it takes so long for us to embed AI

or technology into our companies and everyday work life. So, uh, this slide is very busy. I did this on purpose because you probably see corporations always put out these strategy slides. It usually makes no sense. It's very very hard to translate strategy into what does it actually mean for me as a as a colleague. Um so I actually had to cut this slide down for today's presentation.

So it was that busy. But what I wanted to just highlight to you is that for large corporations, you probably just heard it in the last few presentations, we do need some sort of return on our investment. So ultimately what we're really seeking is where AI transformation can deliver value for our colleagues, clients, and investors or stakeholders. So that word balance in the middle is the most

critical word for us. If there's no balance to what AI is delivering on both sides, the client or colleague, we just don't really see a point to it, right? So, we have to deliver certain outcomes to both our stakeholders as well as our investors. Ultimately, um my job is to help bring this day-to-day to many of my colleagues as well as clients that we work with. and

I'll show you why it's such an interesting role to be in. Now, Marsh itself is 95,000 colleagues around the Not only do we have to balance our internal client and colleague needs but we also have to advise thousands of corporations as to how they should look at into AI and the risk they will face uh the compliance uh and regulation that they will face as well as

the workforce design and optimization that they will have to uh deal with. So, not only are we having to do this for ourselves, but we simultaneously consult uh many other corporations. One of the first things that um I had to deal with, of course, is the EU AI act. And I thought it was very relevant because we are based in Europe. Um as an American, it's very

frustrating when I have to deal with EU regulation because it's the highest regulated region. So, it makes my job very entertaining when I have to read. so much law and legal documentation, but I'm lucky that I have a whole legal team that helps me get through that. But, uh, it's the number one thing that we're warning our clients about. Make sure that you understand the laws and

where you're operating, where you're sending your products and services. Ultimately the what we're also designing at the same time trying to understand the laws to make sure we're compliant is why it takes us so long to implement AI [clears throat] currently today. So we do have our basic foundational AI. We also buy very very customized AI products that make our staff faster and leaner. Um, but we

have to have products ready and consulting ready and expertise for our clients so that we can tell you where you need to watch out for risks. So the EU AI act, you heard about it earlier. Please make sure you don't adopt AI or you don't start changing your workflows until you really really get the right advice. Um, on the slide I also had the AI and cyber

uh literacy. So what does that mean? It means that we have to ensure that 95,000 people know how to use AI products. They have to know what cyber security is, what confidential information you should or should not be putting on any of these tools that we have. It's a it's a huge workload. We also make sure that we have the right data access throughout the company. We

don't give licenses to every AI product out there because it's uh extremely confidential information that we work with and yes it takes us time typically as part of a person I co-chair a risk committee in Europe it takes us about six months to review a new product or service that we need to offer to clients if it has AI as a component. Yes, you heard me. six

months because it goes through so many checks and balances for us to make sure that we're offering our clients something that is compliant and safe and secure. That means we also have to train thousands of colleagues typically on that. And then of course we have to uh make sure that our clients are ready to do the same for their own colleagues, right? So we do have human

capital consulting products. We have cyber um literacy surveys, if you will, where we can actually um have your colleagues take a cyber security literacy service survey, excuse me, and we can tell what is the risk of that person actually uh clicking on any fishing, any scams they receive. So, we know when your staff uh behaviorally may have some um challenges and risks that they can bring to

your organization. And that could lead to of course more uh training for them just to be a lot more literate and aware of what's out there. Uh deep fakes, you've heard it all. So we do do that type of consulting and training for our own staff as well as many uh clients out there. [snorts] And then um what I will do is just touch upon workforce design

and optimization. So you heard it before on the stage is everyone's always worried about is AI coming for my job? So typically no, we don't see major uh layoffs coming. We don't typically do that in our company either. It's really meant to restructure what is it that you're going to do daytoday. That's what I do with my own team. They use AI every single day. They're also

younger than me. I'm the boomer is what they call me. um on the team. So they are far more advanced with using AI when it comes to strategy than I think I am today. But ultimately it's really as a manager for me it's shifting the tasks that I can assign and the timelines and the deadlines because they're just much faster than we have ever been before. So

let me touch upon [snorts] where we are as a company today when it comes to how we're going to redesign our workforce. So on this next slide where you're seeing is where we are as a company. We already do have the foundational AI. So chat GPT cloud they operate in secure environments. Everyone can use them. They can access our files, emails. It's all at their disposal. But

where we're moving to next is that aentic AI, which you also heard others touch upon today. That's where we no longer are interested in scale. To be honest with you, we don't need any large language models across 95,000 colleagues. What I look for and what my leadership team will look for is very very specific products. So I need a solution that very tangibly changes somebody's uh workflow.

It needs to answer very very specific questions. So we do buy very specific vendor AI products and capabilities that have helped us um already have AI that can make decisions instead of uh colleagues. But those colleagues we've redeployed to something else. They're usually experts who can continue consulting clients. They're no longer keying. They're no longer doing data entry. They're not checking on insurance policy and seeing whether

that's uh been covered for a client. That the AI can do that for us and we've purchased that and we will try to scale it. So the future of course I can't tell you what that will bring but I can see what's happening today and this is where we have a lot of client interest and of course colleague interest as to how their dayto-day will change. I

personally look at how can I restructure teams for the better. It doesn't mean I'm trying to get rid of human beings because I don't think that judgment will come from AI. Hence why we're just trying to apply the judgment where we can and see and test out and pilot different products. But again, it's highly specific cases that we're trying to test out at the moment. So in

the last few minutes, what I thought would be interesting for you as well, uh, as I mentioned, we're the largest risk advisor in the world. So what we do is we are effectively an insurance broker for the largest corporations out there. We typically don't uh work with uh individuals. So if anybody has car insurance issues, please don't come to me. Um it's I I only deal with

large cyber security issues. We have financial risks that we deal with uh and find insurance for. One of the biggest things we've had to create this year and launch in 2026 as uh Marsh especially Marsh Europe is we've uh launched our digital infrastructure team. So that is a highly skilled team of people who are experts in the field of digital infrastructure. So what does that really mean?

It means that we have to consult our clients from the nent or the initial project of wanting to build a data center having that energy or electricity that you heard about earlier that's required to really give you that access to AI and the power it it needs. So data access data centers and regulators it's that life cycle from putting a project together to constructing a data center

uh making it operational and [snorts] then actually uh we also work on decommissioning data centers that are no longer fit. So data access itself you heard it earlier really really challenging. uh we still as I mentioned we uh advise 16 trillion dollars worth uh for other co corporations data access is still one of the key reasons why they do not implement AI so an AI is not

making a decision on your pension it's not making decision on your investment we just don't have they just don't trust that they have the right data the data centers I just spoke we offer um consulting services and risk services across the entire life cycle But one of the biggest concerns I have is that a high percentage of the global data centers are built in climate risk zones.

So imagine a flooding, hurricanes, tornadoes. If they hit those data centers, there will be no power in them. What is your business going to do when you have that business interruption? It's something I think about because if we become reliant on it, it is a scary thought as to how quickly we can actually get power back to them. And uh business interruption is something that we personally

advise uh many clients on. And lastly, regulators. So regulation EU AI act, I'll mention it again. It is going to most likely become the baseline for the entire global AI regulation. That's what I'm seeing is being predicted. It is the strongest law at the moment. It's not necessarily going to go away. So my hope for you is that you do seek the advice once again because the

regulation is here to stay and most likely will continue coming as we advance through AI and the products and services for clients and for ourselves. So, thank you so much for your time and please uh come and speak to me anytime you wish. Thank you.