Marios Rotsides: For us at Wargaming, AI isn’t plug-and-play

Large Language Models like GPT aren’t products: they’re enabling technologies. And to make them useful, especially in a complex gaming ecosystem like Wargaming's, the company had to integrate them into existing frameworks, fine-tune them and rigorously test their functionality, according to Marios Rotsides, AI Technology Product Manager, Wargaming.

In an interview in GOLD's May issue, Rotsides also reveals - among other - what is the one AI breakthrough that he believes will be a game-changer for the industry.

What are the specific AI tools that you have adopted? How well did they integrate with your existing infrastructure? Did they require specific customisation to work effectively?

When most people think of AI today, they picture ChatGPT, Tesla’s self-driving technology or virtual assistants like Alexa. AI is the buzzword of the decade – equal parts hype and hope, fuelled by FOMO (fear of missing out) and significant investments. For us at Wargaming, AI isn’t plug-and-play. Large Language Models (LLMs) like GPT aren’t products: they’re enabling technologies. To make them useful, especially in our complex gaming ecosystem, we’ve had to integrate them into existing frameworks, fine-tune them and rigorously test their functionality. While we’ve adopted some smaller third-party AI tools for daily workflows, production-level integration requires much more groundwork.

Have these AI deployments led to revenue growth, cost savings or improved customer retention? Can you quantify these improvements? And how long will it take before you see a measurable return on your investment?

Big corporations like ours take time to adapt to disruptive technologies. The single biggest business in 2024 was probably for consultants explaining what Gen AI is. In 2023/24, many companies made significant upfront investments in AI but saw little immediate return. It’s a long game. Just as the internet took over two decades to achieve mass adoption, AI needs time to mature and find its product-market fit. Only 20% of enterprise use cases made it to production status in relevant industries like software development, while industries like Legal, HR and Sales were well below 10%. We’ve been blending AI with machine learning, data science, and statistical modelling for over a decade but with generative AI (Gen AI), we didn’t dive in headfirst. Instead, we dipped our toes in it with research and small-scale experiments before committing to larger investments about 18 months ago. We are laser-focused when it comes to measuring impact – it’s part and parcel of developing solutions. This approach allows us to objectively assess whether what we’re doing is making a difference. As the first year or so was more about testing the waters, our current solutions now need to demonstrate a tangible impact on revenue.

What have been the biggest bottlenecks in integrating AI solutions? How did you overcome these challenges?

Our main challenge has been maintaining the high quality of our flagship games, like World of Tanks and World of Warships, while introducing new technologies without compromising the player experience. It’s easier to experiment with something new than to risk disrupting established products played by millions daily. For Wargaming, which has always embraced cutting-edge tech, we’ve taken a cautious approach: calculated risks, thorough testing and ensuring that any integration aligns with our commitment to quality.

As AI continues to transform the workplace, how has Wargaming adapted? How have employees responded and what steps have you taken to support this transition?

AI hasn’t sparked a workplace apocalypse yet. It’s unavoidable in tech-driven industries like ours. Employees realise that AI isn’t here to replace them but to enhance their work. The real fear isn’t losing jobs to AI; it’s falling behind peers who embrace it as a tool for growth. We’re still at an early stage of this journey but the mindset is shifting: using AI effectively is becoming a competitive advantage for individuals and teams alike.

In your industry, what types of decisions can realistically be entrusted to an algorithm? And do you think that AI will ever be capable of handling the novelty and ambiguity of executive-level decision-making?

We remain a human-driven company where player satisfaction is paramount. Right now, AI excels at assisting humans, whether by speeding up development cycles or processing massive datasets beyond human capacity. As for executive-level decision-making? Never say never! But today’s AI lacks the sophistication and trust needed for autonomous leadership decisions. It can support decision-making but not replace human judgement.

Which AI trend or technology are you watching closely and which do you think are more marketing fluff than substance?

AI in healthcare is fascinating. Predictive analytics, drug discovery and personalised medicine are reshaping the field. Agentic systems capable of autonomous decision-making are also gaining traction in customer support and enterprise operations. On the other hand, AGI (Artificial General Intelligence) remains a distant dream. Progress has been slower than anticipated and the gap between narrow task-specific AI and AGI is still enormous. Generative AI has been both exciting and overpromised. While it’s great for content creation – text, images, videos – it hasn’t replaced human creativity as some predicted. The human touch remains irreplaceable in originality-driven fields.

What is the one AI breakthrough that you believe will be a game-changer for your industry?

Personalised gameplay will revolutionise gaming. Imagine experiences dynamically tailored to each player’s behaviour, preferences and skill level – making every session feel fresh and engaging. Features like adaptive storytelling, intelligent NPCs (non-player characters), dynamic difficulty adjustments and personalised content recommendations could redefine how players interact with games in the near future.

• This article was first published in the May issue of GOLD magazine. Click here to view it