Reading Time: 3 minutes

“Movers and Shakers” is a dynamic monthly column dedicated to exploring the latest trends, developments, and influential voices in the iGaming industry. Powered by GameOn and supported by HIPTHER, this op-ed series delves into the key players, emerging technologies, and regulatory changes shaping the future of online gaming. Each month, industry experts offer their insights and perspectives, providing readers with in-depth analysis and thought-provoking commentary on what’s driving the iGaming world forward. Whether you’re a seasoned professional or new to the scene, “Movers and Shakers” is your go-to source for staying ahead in the rapidly evolving iGaming landscape. 

By Claudia Heiling, Co-Founder & COO, Golden Whale

For years, iGaming has considered itself a data-driven industry. We’ve all spent time refining segmentation, optimising CRM journeys, mapping behavioural signals, and building increasingly complex player models. And with machine learning now widely available, whether bought, built, or borrowed, it would be reasonable to assume that the industry is already fully realising the benefits of AI.

But speak to most operators, product teams, or data leads and you’ll hear a different story.

There are models running somewhere – and usually several. There are predictions being generated. There are dashboards, reports, and insights circulating. Yet the business impact often feels inconsistent. Some initiatives deliver a clear uplift; others stall or never make it past a proof-of-concept stage. Projects that shine in testing environments don’t always translate into live, reliable operations.

The issue is rarely the model. And it’s rarely the data team. The gap is operational.

It’s one thing to build machine learning models. It’s another to make them function as part of the daily working rhythm of an iGaming business.

The operators and providers seeing the strongest and most reliable gains are the ones who treat AI not as an experiment, but as a capability: something that must be designed, deployed, monitored, re-trained, and continuously improved. This is closer to how we already treat core game operations, promotional systems, risk tooling, or CRM orchestration. It’s iterative, structured and ongoing.

In practice, that means building the frameworks around the models, not just the models themselves. Continuous data flows. Automated re-training. Real-time deployment pipelines. Feedback loops that allow systems to learn not just once, but constantly. When we work with iGaming clients who have embraced this operational mindset and leverage our ready-to-deploy MLOps system built for iGaming, the impact becomes both compounding and predictable.

The other shift happening is cultural. There has been a lingering expectation in some corners of the industry that AI will replace manual decision-making entirely and that it will “take over” processes like CRM optimisation, fraud detection, or product adjustment.

That’s neither realistic nor particularly desirable.

iGaming is too contextual, too human, too dependent on craftmanship and intuition.
The real value of AI is in augmentation: giving teams better visibility, faster feedback, and stronger evidence on which to base decisions.

In organisations where this mindset has taken hold, you see a different dynamic.
CRM teams run more experiments, more often, because they aren’t spending time rebuilding segments from scratch. Analysts spend less time on manual spreadsheet simulation and more on strategic exploration. Live-ops managers can respond to player behaviour as it changes, not after the weekly report comes in.

AI becomes the layer that enhances judgement, rather than replaces it.

And when AI is integrated technically and culturally, the commercial outcomes are hard to ignore. In setups where continuous learning pipelines are properly established and aligned with live operations, we’ve seen engagement and retention metrics improve dramatically and sustainably, with activity and revenues rising by 100–200%, while bonus and incentive costs drop by 20%+, driving growth and both securing and expanding market share. Operational teams benefit too, with workflows becoming smoother and less manual because the system is handling the constant data processing and iteration.

The improvements don’t come from having more complex algorithms. They come from having a structure that allows those algorithms to perform reliably, adapt to change, and keep learning over time.

This is where the conversation about AI in iGaming is quietly changing.

It’s no longer dominated by model performance or dataset scale, rather it is focused on repeatability, reliability and learning speed.

The distinction matters because it separates having AI, from running AI.

And the operators and providers who get this right aren’t just improving performance in the short term. They are building organisational momentum, a capability that compounds over time and is very difficult to replicate quickly.

In a sector defined by tight margins, competition and rapidly shifting player expectations, that advantage is significant.

So, if there is a “next step” in the industry’s AI journey, it’s not a more complex algorithm. It’s not a bigger data pool. And it’s not a new suite of predictive dashboards.

It’s the ability to learn continuously, responsibly and at scale.

Because in iGaming, as in intelligence, data alone doesn’t win. What wins is the ability to turn learning into action again and again.

The post Movers and Shakers – From Data to Decisions: What It Really Takes to Make AI Work in iGaming appeared first on European Gaming Industry News.