Beyond the hype: how OpenBet is building an AI-first operating model
As AI moves from experimentation to infrastructure, OpenBet CTO Maria Kollia explains how the company is embedding intelligent systems across trading, player protection and product development, while balancing automation, governance and responsible gambling at scale.
Where is AI delivering tangible value across OpenBet’s products today?
AI is already deeply embedded across several of our internal pillars, and the foundation underpinning all of it is structured, governed adoption. Early on, we made a deliberate move away from ad hoc experimentation toward a model where every AI initiative operates with defined duration, clear success criteria, and accountable teams. That discipline has allowed us to scale adoption at pace, across engineering, QA, support, product, commercial, and back-office functions, without sacrificing the governance required across all of our global hubs.
To give a sense of the scale, over 600 of our engineers are now working with AI tooling as part of their daily workflow, which alone represents a material shift in how we build and iterate. Beyond developer productivity, AI is delivering value in player protection, where our Neccton technology identifies at-risk behaviour in real time, and in precision trading, where AI sharpens the accuracy of our models to protect margins and enhance the player experience. We are also using AI for operational automation: workflow processes that previously required human execution are now handled end-to-end, significantly accelerating our product development lifecycle and getting innovation into the hands of our partners faster.
OpenBet has chosen to work with multiple hyperscalers including Anthropic, Google, Microsoft and AWS. What sits behind that strategy?
Our multi-provider approach is based on four very deliberate factors. First, market volatility: anyone that follows the AI landscape even semi-closely can see how quickly it is fragmenting, not consolidating, and there is a new model emerging every few weeks. It is clear that no single tool will win everywhere. Second, problem diversity: different challenges across our product stack require different capabilities, and a single provider cannot optimally serve all of them. Third, workforce diversity: our engineers have varied backgrounds and tooling preferences, and our stack reflects that. Fourth, evaluation discipline: we assess tools at the ecosystem level, not in isolation. The question we ask is not “is this good?” but “how does it fit our stack?” Together, these ensure that OpenBet and our partners always have access to the best available capabilities as the technology evolves, without being captive to any single vendor’s roadmap.
How does that multi-partner approach help you stay flexible as the technology continues to evolve?
Our strategy six months ago looked different from today. That is by design; we built an approach that evolves with the market rather than one optimised for a single moment in time. We no longer evaluate tools in isolation; we think at the ecosystem level, which means we can integrate new capabilities into our core stack the moment they prove their value. We are also actively diversifying through open-source, which builds optionality and ensures we are never locked into a commercial vendor’s roadmap. The result is an infrastructure that functions as a living system, capable of pivoting quickly toward the most effective breakthroughs while maintaining the stability our partners require.
What are the key developments in the latest iteration of Neccton?
The latest version of Neccton represents a generational leap in how we approach player protection and integrity. The flagship development is the integration of a unique, generative AI-enhanced assistant. This moves the player protection technology beyond standard predictive modelling and pattern recognition, toward a system capable of deeper player analysis and immediate support triggers. By leveraging real-time operator data, Neccton constantly iterates and improves, shifting the paradigm of player safety from reactive monitoring to a more sophisticated, proactive engagement layer that sets a new industry benchmark for responsible innovation and effective risk management.
The new generative RG assistant is a notable use case. How does it work in practice, and what impact do you expect it to have?
In practice, the RG assistant analyses live operator data to identify subtle shifts in player behaviour that may indicate risk. When certain triggers are met, it provides nuanced, conversational support that feels more personalised and relevant to the player than traditional automated messaging. We expect this to have a significant impact on both player longevity and brand trust, as it transforms player protection into a proactive, intelligent interaction layer. By providing immediate assistance tailored to specific behaviours, we move decisively toward a safer ecosystem where technology enhances human expertise, ensuring that intervention is both timely and effective for those who need it most.
With increasing focus on responsible gambling, how are you ensuring AI supports more proactive and effective player protection?
We ensure AI remains a force for good by treating compliance as a foundation of innovation. Our approach is rooted in a human-in-the-loop philosophy, where AI-driven insights augment, rather than replace, human expertise. To maintain accountability, we have established a cross-functional AI Ethics Committee to ensure every output is auditable, explainable, and fully compliant with global regulations. By combining real-time pattern recognition with expert oversight, we turn player protection into a proactive strategy. This ensures that as we scale our automation, we never lose the transparency or nuanced judgment required to protect players responsibly and effectively.
Looking ahead, where do you see the most meaningful opportunities for AI in betting over the next 12–24 months?
AI-driven player protection and automated trading will continue to shift from differentiators to baseline expectations across the industry. The operators who lead will be those who move beyond standalone tools to embed AI into their core operating layer, building integrated, intelligent infrastructure where trading, risk management, and product development feed into one another in real time. The winners will balance high-speed automation with the entertainment value that remains at the heart of the sportsbook experience.
