Leading AI Transformation Without AI Experts: A Conversation with Simo Dragicevic

In its latest report, Amrop’s Global Digital Practice examined the leadership competencies essential for successfully integrating AI into organizations. We invited CEO/GMs from midsize, PE-backed, family-owned and other companies, to share their real experience in leveraging AI strategies for their organization and customers.  

Mikael Norr, Managing Partner at Amrop's Stockholm office, spoke to Simo Dragicevic, who has been an entrepreneur in the gambling industry since 2010, as founder of BetBuddy, a pioneer in data analytics and responsible gambling, which was acquired by Playtech Plc in 2017.

Dragicevic currently sits on the boards of Canada's Responsible Gambling Council and the GB Gambling Commission, as a member of its Digital Advisory Panel. Most recently he has started working on the launch of GSI (www.safer.games), focused on developing new insights at the intersection of products, players and safer gambling.

Simo Dragicevic Interview

Mikael Norr: How do you see the future in using AI tools within the gambling industry?  

Simo Dragicevic: It depends how far you look into the future, but, I guess, if you think about the near-term, the next five years, then I believe the application of AI is going to be around improving productivity across a number of domains in the industry. If you think about the value chain, if you think about the customer, and the interaction with the customer, we will continue to personalize the customer experience. So, putting a lot more personalization and context around the user interface in terms of what people see, what events and gambling offers they see. So that's obviously something that's been happening for some time, and it will continue. The gambling industry is a little bit behind the big tech, so there’s still room for improvement. I absolutely see forms of communication with the customers, especially those of younger generations, being augmented with generative AI chatbots and so on, and customer service representatives will probably be prompted to support their customer interactions similarly. Also, if you look at the generation of content within gambling, there’s sports betting where there’s a potential to automate, as there are 1000s of different sporting events or lines that one can bet on. The trend in sports betting is to produce more markets and more betting opportunities, especially around what is called micro betting where you can bet on the next pitch of a baseball game, for example.  

MN: That must be automated at some level already, right? Or is it supervised by people?   

SD: If you look at what the newer sports betting companies are doing in the last few years, a lot of that is automated, but I believe there is still human oversight. Operationally I don't know when a human steps in, especially with these shorter timeframes, but you can absolutely see how that could be automated even more and how more automation will create more markets and opportunities and more variety. And if you think about other big gambling verticals, like online casino slots, table games and so on, you can absolutely see how generative AI can help the creative process of developing content and images, making it easier to develop new forms, themes and content.  

MN: Let’s think for a moment about disruption. For example, in my industry, I get calls every week from companies saying that they can help us with our processes, they can automate everything we can think of and so on. But, probably within the next two years it’s only a matter of increasing productivity a bit, it’s not disruption. Do you see any disruptive effects of AI in your market?  

SD: We talked about micro betting within the sports betting industry – the ability to automate it consistently produces lots of new betting opportunities. In the US there’s a very popular form of betting - parlay betting, which we call an accumulator, where you can combine events and build them up. And being able to offer these interesting combinations consistently, while the game is happening, having that level of automation is somewhat disruptive, because it continues to offer more and more ways to make the game interesting. So, it’s not a completely new product, rather a variation of an existing product, but I believe there is an element of disruption here.

Another area where you could get some interesting opportunities is the casino furcation of sports betting where you can effectively build a slot machine around the events that are happening in the game – basically adding another level of enjoyment and thrill through this new gaming format. The casino mechanics could lend themselves to sports betting, so I believe this can give rise to innovation bordering on disruption. 

MN: I have just one more gambling-related question: is the “back end” of all the payment solutions ready for this type of micro-gaming activity? If there’s, for example, 100 bets placed in a round of golf and if you treat them the normal way with various small transactions, couldn’t that cause such a volume that it wouldn’t be profitable for the company?  

SD: If you look at the financial markets, this very low latency, the immediacy of information, the large volume of trading – the infrastructure could be built. And it is a genuinely good question around AI, because people underestimate the infrastructure required – not only the processes and run, but also the need to train models in advance, the data and processing requirements. At the same time, you want to be reasonably certain there’s some interest in the product so you would probably start pushing it and see how it goes, but there you have the actual source of the data, and the data provider has to push it to the odds generator, which then has to push it to the book who then has to push it to the person who has to make the decision so that there may be some limitations in how far you can push it.  

"When you don’t have anyone with Data, Machine Learning or AI experience, there’s a nervousness around tackling it in your own organization."

MN: It is generally an interesting topic – about the amount of data. Take automated driving, where you could have 10,000 cars on M1 communicating with one another, which means thousands and thousands of data points, which you need to handle in some way; you cannot store it in the cloud – there needs to be another type of direct communication. Of course, this is something you would probably also use AI to handle, but there needs to also be capacity and infrastructure. 

SD: Absolutely. When you think of the costs for just training ChatGPT you start understanding what goes into these models and realize why Open AI had to have Microsoft as a partner. Because to have this computational power, to have the GPU capacity, we’re talking tens, even hundreds of millions of dollars. There’s this debate around big tech being the only ones who can win this race because they’re the ones who have the capacity and the funds to build these models… 

MN: Thinking of what we’ve just discussed and looking at a typical leadership team, about your last leadership team or even the non-executive board of directors that you sit in, what type of competencies do you see that you’ll be needing in the future that are perhaps lacking now? Is there a need for specialist competencies to deal with AI-related matters? 

SD: What I’ve seen is that when you don’t have anyone who has tech, data, machine learning or AI experience, there’s a nervousness about trying to approach and tackle it in your own organization. A recent example for me is the UK Regulator where I have an advisory role, and which, as a quasi-governmental organization is very analog, and a lot of people have been in their roles for very many years. There was an understanding that they’re missing a huge opportunity in terms of doing their job better with the help of data, so a couple of years ago they said that they needed a data strategy and to modernize, to be a modern regulator. The default was to bring in consultants to help them think through the problems and to then hire an executive who will lead the new program. Now, I think, on both fronts it was probably the wrong thing to do. Consultants can be very useful in facilitating progress, kickstarting things, but there wasn’t enough curiosity from the executives – it was more of: let’s get some consultancy in and give them the task to help us figure this out, rather than: what do we actually want to do here and how do we want to change ourselves, how are we going to make gambling fairer, safer and crime-free through the use of tech and data? 

MN: What was your advice in this case? 

SD: My advice was to find someone on the executive board who is going to carry the can for this. It’s probably best that it’s the most digitally native person who’s the most comfortable with tech and data, but this needs to be part of their day-to-day job, their mandate. They need to free up their time, so they can focus on this and then slowly, working with advisors, whether it’s myself or the other panel members, they begin to build a strategy, understand what skills are needed and start building that way. And what’s needed first is curiosity at the executive level, asking the right questions, accepting one’s limitations, admitting what scares you, understanding what doesn’t, where we’re strong, and getting support in the areas where it’s required. So, I don’t necessarily think that it's right to “inject” someone in the organization who’s going to solve the problem – working internally might take a little longer but I think it sets the organization up for success in a much more sustainable manner. 

MN: So, the key word is curiosity. 

SD: Absolutely. I was reading the annual letter of Jamie Dimon, the CEO of JPMorgan Chase, where he talked a lot about AI, and they’ve created a new role – the Chief Data and Analytics officer, but they’ve put a seasoned executive in that role, who’s been with the organization for 20-30 years. I know that they also bring in a lot of experts and they hired a head of AI many years ago already, but they’re not the people who make the decisions, set the strategy and run the organization – they’re the people who support the executive in realizing their vision. I see that as a good template where technology becomes the last layer in the jigsaw, and that’s what we’re trying to do with the Commission as well – very different organizations but similar strategies and philosophies when it comes to AI. 

MN: In your experience, are the CTOs/CIOs in a typical organization tend to be more pro or against the extensive use of AI? 

SD: I think it depends on the individual and yes, there will be some people who are against it. When you’re in a tech company the CTO needs to think about keeping the wheels running, about 24/7 real-time support to the system, so maybe asking them about the architecture – how do we bring data in and make it accessible to business analysts and data scientists in four or five different parts of the organization, might not yield the best results. But at the same time, they’re fundamentally critical in helping to enable that vision. At the same time a lot of the CTOs I’ve interfaced with tend to want to use the latest technology – they want efficiency; they want their teams to be learning and growing. In my personal experience at Playtech (a leading gambling software development company) – it was a very complex organization, it was kind of like moving a tanker: there were lots of legacy platforms and tech, and we ended up putting layers and layers of databases and technology to try and knit it together, which was probably not the most efficient way of doing things, but demonstrates that it can really be very hard. 

MN: During your time at Playtech did you launch any AI tools internally that you could measure the success of?  

SD: With Playtech we were moving towards business intelligence reporting and analytics, trying to embed that more widely within the organization; then there were pockets of machine learning in different parts of the business, but they tended to be isolated. There was more focus on what I call BI and analytics tooling to make better use of data to support business decisions, and I think that a lot of the gaming industry is probably in that position.  

MN: You were in the recruitment process for the role of the CEO at Kambi, the B2B provider of sports betting services to licensed B2C gambling operators. What would you do at Kambi to get people in the management team to the next level when it comes to the use of AI? 

SD: When you come in fresh and new into an organization, you want to have a good look across the different areas and departments, to understand the key processes – what happens in sales, customer support, the engineering team and the sportsbook team, what happens in operations. And then it’s also good to look outside of the organization, bring in examples of how other organizations have done things, how larger tech companies organize themselves, how their teams are made up. The way I would approach it is - let’s understand what we do today, who we have, how we’re making this work, and how we can improve it. Do we, for example, want lots of junior traders pricing all these markets, do we see AI dominance in these processes, in software development, do we need all these testers and engineers? Ultimately, it’s a people organization, so how can we make people’s jobs more meaningful; how can we free them up to do more interesting, more value-added work, to sell more effectively, and get better insights from the data? So, it’s really the challenge of trying to really understand the organization. 

MN: It’s interesting because it comes down so much to the personality of a leader who may or may not be reluctant to open up, take input from outside without trying to protect the legacy or the people already working there, because ultimately, it’s the company objective that one should have. 

SD: Absolutely. And when an external person comes in, they can be nervous, because as an external you can ask really dumb questions, really simple questions, like, how many people do you have, what do they do, why do you have five people in this team doing this or that? You can ask these questions because you don’t have a legacy, but also you have to be careful with that approach – it’s not to catch people out, but to help me understand what they’re doing and where the problems are, you can bounce ideas off me. And it’s a good question: should you have an AI division? One of the things that I’ve always done is try and bring in academics. It’s interesting to work with them because they’re very deep in their topic, and you can tap their knowledge, their networks, their ideas, and that can open up a whole realm of possibility that didn’t exist before – because they’ve spent 20 years researching computer science and AI, and bringing this expertise together with the experience within the company can be a great discovery process. 

MN: They might also feel less, let’s say, threatening to the people within the company, compared to people from a consultancy firm? 

SD: I’ve been a consultant myself and I know the value they bring in the right environment, but I think in the case of engaging academics you’re basically trying to extend the brain of the organization. If you look at Microsoft, Google or Meta, they’ve been tapping into the best academic brains for decades – and it’s no coincidence how successful they are in this space because it’s a constant flow of ideas. And then it’s a kind of symbiotic process – you bring the academics in, you give them data, you give them space, you let them publish, you let them work on big, exciting problems. And over time you can really create some snowball effects: you don’t need a whole computer science team from Stanford – it can just be one academic, or it can be a team of five people, and you give them space, and the environment and data and some resources, and some small, immediate goals. And then you have a demo, you have a conference paper, you have an industry trade press article, and you compound that over several months and years, and all of a sudden, you’ve created something that’s actually quite hard to replicate. 

MN: And this is something you can do on your next journey as well! From what I hear, also, it’s not a very often-used approach, not as often as it should be. 

SD: I don’t think it is, and again, maybe it comes down to curiosity and experience. The gambling industry is an interesting one: you have to look at the history of the industry. It was created not that long ago, and some of the entrepreneurs made a lot of money from not really doing that much - putting up a casino site and collecting the deposits, as there wasn’t really much science behind it – and that’s a bit of a disservice. These people, the companies grew, and then you have a corpus of people who probably never once thought about trying to solve the problem from a different perspective. But it’s changing and the Scandinavian entrepreneurs, as I see it, are different and more socially responsible. 

MN: One of the questions we wanted to ask is about where you expect to hire the people who would take care of the AI initiatives from? So, even if you don’t hire, you would be looking more towards academia, when it comes to additional help with it? 

SD: Companies acquire talent all the time, and there are companies which will acquire AI companies because it’s quicker to acquire them. It happened to my company, BetBuddy (a pioneer in the use of software and AI to protect consumers from gambling-related harm), because it would just take a long time to build what we’d built, and there’s time and place also for that. But I’d like to try and foster that curiosity, discovery, the testing and the development of the internal capability. Unless you try to build that within your organization, I think you’re doomed to fail in this space in the long run. Bringing in pockets of expertise, building long-standing relationships and collaborations with universities and other organizations – these are things that can really help you build capability over time, and I think the organization, the employees will find it very rewarding to be exposed to this kind of opportunity. You have to try to do it from within – you won’t have all these people and all these mindsets, but I believe you can train the executives to be open to these approaches and create the space and opportunities to do that. 

MN: Can the external help, in your experience, be useful when it comes also to facilitating the process of employees working alongside AI, so to speak? Kambi offers a good example for this kind of situation, because they have 200 people working on setting odds and next to them sits an independent company working on automating everything they do. How to get those 200 people into using the new technology, doing something else when the process is done? 

SD: I see it not as a revolution, but an evolution, so the way I would position it is: why are we here? If we want to grow as a business, and if you want to advance personally, we need to work differently. It doesn’t mean we are all out of a job, it means that we need to do better work, smarter work, to change our skill sets and rely more on technology. I would ask the people what they want to be doing in five years’ time, make them be part of the decision about how to evolve their roles. You have to lay out a vision – that they can be part of this, benefit from this trend that’s happening, but it’s not going to happen if we all cannot face some uncomfortable truths – that we need to change, adapt and embrace this. I would always try to frame it as a broader vision that everyone can relate to and aspire to. 

MN: And these bilateral discussions generally need to end, there needs to be a group vision about where we’re going. When we talk about AI, if you are on the executive board, you need to be informed about what happens on the other side of the company. 

Simo Dragicevic: Yes, and there’s often lots of different things happening in different silos that no one’s aware of. What you want instead is to get everyone’s heads together in a safe, non-critical environment, looking at the facts. Transparency is really important, and you need to understand what’s happening in the organization in order to drive it forward. 

Key Takeaways

Effective integration of AI requires curious, digital native leaders who actively seek to understand and explore AI opportunities. Nurturing internal expertise rather than relying solely on consultants is vital for sustainable success, including collaborations with academia.

In the near term, AI in gambling will enhance efficiency, personalize customer interactions, and automate content creation. While not entirely disruptive, innovations like micro-betting and new gaming formats will create evolving opportunities that may approach disruption through increased market variety and engagement.

Successful AI adoption depends on breaking down silos and involving employees in the transition. Framing AI as an evolution - focused on smarter work and growth - encourages acceptance and cross-functional collaboration, and helps embed AI into the company's strategic vision.

To find out more, reach out to Mikael Norr or the Global Digital Practice members in your country.