AWS Global Games Industry Business Development lead Shayan Sanyal outlined the technology and data strategies operators can use to improve player experiences while managing scale, personalization and compliance during the Player Experience Academy at SBC Summit Lisbon.
Held on Tuesday, September 29, the half-day programme was part of SBC Summit’s Education+ track and targeted senior decision-makers with practical frameworks around engagement, retention and compliance. The session took place at Sala Tejo at MEO Arena.
Sanyal argued that operators face a twin challenge: handling large fluctuations in demand while delivering experiences that feel increasingly personal to individual players.
“Games and betting are among the only products in the world that have to earn the user’s attention every single second,” Sanyal said, highlighting the pressure on operators to deliver compelling experiences from the beginning of a player’s session.
He said scale and personalization should not be treated as separate engineering challenges, but as two outcomes built on the same foundations of cloud infrastructure, data and machine learning.
From efficiency to insight
Sanyal pointed to Soft2Bet as an example of how cloud migration can provide an initial efficiency benefit while also creating the infrastructure for more advanced capabilities.
According to AWS, Soft2Bet’s migration to AWS reduced compute costs by 55%, improved time to market by 200% and reduced the time required to onboard new partners by 70%.
Sanyal said the more significant benefit comes from what operators can build once their data and infrastructure are more flexible.
He highlighted Soft2Bet Chief Product Officer Yoel Zuckerberg’s comments on the company’s previous reliance on multiple on-premises data sources. Moving to AWS allowed the company to establish a single source of truth for real-time or near-real-time analytics, while adding new data sources became a configuration exercise rather than a hardware project.
For Sanyal, this illustrates the progression from efficiency to insight. “Efficiency gets you onto the cloud, but insight is really where you decide to stay,” he said.
Building for peak demand
The discussion then turned to scalability, particularly the challenges created by major sporting events.
Sanyal said traditional infrastructure planning often requires operators to buy capacity for peak demand, leaving resources underused outside major events. Elastic cloud infrastructure, by contrast, allows operators to scale capacity up when demand increases and reduce it once the peak has passed.
He cited SPRIBE and its Aviator crash game as an example. The company experienced rapid user growth while supplying nearly 5,000 casinos across more than 100 countries.
According to Sanyal, moving the infrastructure to AWS allowed the company to increase its peak handling capacity and process four times more bets per minute while reducing operating costs.
He also highlighted Betsson’s approach to operating across multiple regulated markets. The global operator runs more than 20 brands across Europe, the Americas and Asia, with each market bringing its own regulatory requirements.
Sanyal said Betsson built a global network on AWS designed to put infrastructure closer to players while meeting local regulatory requirements, with the operator continuing to work on security and latency across its operations.
Personalization goes beyond segmentation
Sanyal then challenged the industry’s approach to personalization, arguing that much of what is marketed as personalized experiences is actually segmentation.
Players are typically placed into broad categories such as high-value customers, new players, sports bettors or casino players, after which they receive experiences based on what has worked for others in the same group. “It’s useful. It’s better than nothing. But it is not personal,” he said.
True personalization, Sanyal argued, needs to be contextual, adaptive and continuous. Instead of relying only on historical activity, systems should respond to what a player is doing in the current session, including their pace and behavior. That could influence recommendations, the timing of an offer, the tone of a message or even the way an interface is presented.
AWS has developed guidance architectures for player engagement that use first-party data and machine learning to support recommendations and tailored offers.
Sanyal described this as a flywheel in which first-party player data is used to train models, the resulting personalization generates new player responses, and those responses feed back into the data used to improve the models.
He argued that the quality of an operator’s first-party data is therefore an important differentiator as more companies adopt similar AI and machine learning technologies.
Engagement and player protection
The session also addressed the relationship between player engagement and responsible gaming.
Sanyal rejected the idea that the two objectives necessarily have to work against each other, arguing that personalization and player protection can rely on many of the same underlying data and machine learning capabilities.
AWS has published predictive guidance architectures using Amazon SageMaker that operators can use to train models on their own data and identify potentially risky behavior in near real time.
“Personalizing and protecting are not two separate investments. They are two applications of the same data and the same models,” Sanyal said.
He argued that operators able to integrate these capabilities can build experiences that are both engaging and responsible, rather than treating player protection as a separate technological function.
AI and reducing complexity
The final part of the session focused on how AI can change the way operators build and develop products.
Sanyal cited Entain’s modernization of its proprietary on-premises platform using AWS services, as well as the company’s efforts to upskill engineers through a generative AI hackathon involving 13 offices and 250 teams.
According to Sanyal, the initiative helped Entain reduce the deployment time for new applications and features from weeks to less than two hours. Its software simulation models also saw critical components running in under 450 milliseconds, compared with 2.5 seconds previously.
The broader objective, he said, is not simply to add AI tools, but to reduce complexity and the cognitive load placed on engineering teams. That could allow companies to run more experiments and take more shots at developing new products and experiences without increasing development resources at the same rate.
Looking ahead, Sanyal identified three areas that could shape the next phase of the industry: autonomous and agentic AI, regulatory portability and the possibility of players becoming creators.
He argued that operators will increasingly need to determine which decisions can be delegated to AI and which should remain under human control, particularly as autonomous systems become capable of acting on behalf of businesses.
Regulatory portability will also become increasingly important as markets continue to introduce, change, or reverse regulations, requiring technology platforms to adapt quickly to new requirements.
Finally, Sanyal pointed to the video games industry, where platforms such as Roblox and Fortnite have increasingly blurred the line between player and creator, as a potential model for future gaming experiences.
He concluded that the next phase of competitive advantage will be less about individual campaigns or tactical decisions and more about whether an operator’s architecture, data, organization, and regulatory posture can adapt faster than the market.
“The companies that invest in adaptability today will feel the benefit for years,” Sanyal said.