From Technology to the Experience: Shayan Sanal looks ahead

The ability to understand individual player behaviour is becoming more sophisticated, raising expectations around personalisation, player protection and the quality of the experience itself,  Shayan Sanyal has watched that evolution from both sides of the technology equation. Before joining Amazon Web Services, he spent many years building startups where infrastructure decisions were measured against runway, speed and the practical demands of getting a product into the customers’ hands. Today, as AWS’s Global Games Industry Business Development Leader, his perspective spans games, betting and the technologies increasingly connecting the two.

In this conversation with us, Sanyal examines where genuine technological advantage now lies, why businesses risk being overwhelmed by the pace of innovation and what games can teach the wider digital economy about earning, rather than assuming consumer attention. He also considers a more fundamental shift, from personalisation designed primarily to optimise commercial outcomes towards technology capable of serving the individual player.

The implication is that the next phase of competition may be determined by those who understand where, when and why to use technology. 

Q: You spent close to a decade building and operating startups before joining AWS, initially working with founders and then moving into games technology. How did being on the other side of the table , having to build, sell and scale, prepare you for your role at AWS?

Shayan Sanyal: Honestly, it taught me more than I expected at the time. When you are a small team trying to ship a product, find customers, and keep the lights on, you learn very quickly what matters and what does not. Every one of my startups was built natively on AWS, not because I had a relationship with the company, but because we needed infrastructure that let us move fast and only pay for what we used. We benefited from programs like AWS Activate, which gave us credits to experiment without burning through our runway, and access to solutions architects who helped us make better technical decisions early on. Speed to market was everything, and that support made a real difference. It gave me a deep appreciation for how foundational technology choices, and the network of support around them -- either accelerate a business or slow it down.

At AWS, I now lead a global team focused on games and betting & gaming technology. Having been on the other side of the table helps me listen differently. When someone tells me they have three engineers and a deadline that was already yesterday, I do not have to imagine what that feels like. It also means I tend to focus conversations on what will actually move the needle for a business, rather than what is technically impressive but not yet practical.

The biggest lesson from startups, though, is that building a great product is necessary but nowhere near sufficient. You also need distribution, timing, and the discipline to let go of things that are not working. That mindset translates directly to how I think about the games and betting industry today: the companies that are thriving are not necessarily the ones with the most advanced technology, but the ones that are honest about what their players actually want and how quickly they can deliver it.

Q: Technology is developing at extraordinary speed, with AI, cloud, advanced computing and increasingly autonomous systems evolving and converging at once. Is there a danger that the pace and complexity of change is becoming greater than businesses' ability to absorb it?

Shayan Sanyal: There is definitely a danger, and it is not theoretical. It is happening right now. I see it across every market I cover. Studios and operators are being asked to evaluate generative AI, agentic architectures, real-time personalisation engines, new compute paradigms , all at once, all while running a live business.

The issue is not the technology itself. The issue is the organisational capacity to absorb it. Adoption requires more than procurement. It requires changing workflows, retraining teams, rethinking metrics. Most businesses are still structured to evaluate one major technology shift at a time, and we are asking them to evaluate several simultaneously.

It is worth remembering that AI today is still very early. In many ways, it is where the internet was before cloud computing made it accessible. There are few reusable building blocks. Expertise is concentrated and difficult to access. Every integration point is expensive. Organisations are essentially building from scratch, and the cost of experimentation remains high. The companies handling this best share a common trait: they resist the temptation to adopt everything at once. They pick one or two areas where the technology can solve a real, measurable problem today - fraud detection, player segmentation, build pipeline automation,  and they go deep on those. SPRIBE is a good example: they focused specifically on scaling their infrastructure to handle explosive growth in their Aviator game, and ended up handling four times more bets per minute while reducing costs. They did not try to reinvent everything at once. They treat new technology the same way good game designers treat new mechanics: introduce one at a time, measure the impact, then layer on complexity.

The other thing these companies get right is AI governance. When AI is being embedded into player-facing products, compliance workflows, and operational decision-making, you need a framework for how it is evaluated, monitored, and corrected. Not a 200-page policy document that sits on a shelf, a living set of principles that tells teams what they can deploy, how they measure its impact, and when they pull it back. Without that, the pace of adoption becomes a liability rather than an advantage.

What I would caution against is the opposite extreme, waiting until the technology stabilises. In a competitive market, the cost of inaction often exceeds the cost of a controlled experiment that does not work out. The answer is not to slow down. It is to adapt with discipline.

Q: How do you distinguish emerging technologies that will genuinely change an industry from those that are simply generating noise?

Shayan Sanyal: There is a test I keep coming back to: does the technology solve a problem that people are already spending significant time or money on, or does it create a category that nobody asked for?

Most genuine transformations begin as efficiency plays. Cloud computing started as a way to avoid buying servers. Machine learning in games started as a way to reduce QA cycles. Personalisation in betting started as a way to improve offer conversion. In each case, the value was measurable from day one.

The noise tends to cluster around technologies that require the customer to completely reinvent themselves before seeing any return. If someone has to change their org chart, retrain 60 percent of their team, and redesign their core product just to pilot the technology, it is probably not ready -- or the technology may not yet be ready for production at scale. Compare that to what NSUS Group, the company behind GGPoker, did with machine learning: they had a manual, labour-intensive fraud prevention process that could not scale with their transaction volumes. They built an ML-powered detection system using Amazon SageMaker in three months went from 270 features down to 80, achieved a 95 percent AUC reliability score, and cut their internal management costs by over 80 percent. That is an emerging technology solving a concrete, urgent problem in the betting and gaming space, with results you can measure immediately.

I also pay attention to whether the builders themselves are using the technology. In games, as studios adopt AI coding assistants at scale, or deploy ML models into live production environments rather than sandboxes, that is a meaningful signal. That is adoption at the operational layer, not the PR layer. When adoption stays at the announcement stage without moving into production, it is worth looking more closely at the underlying value.

Q: AWS works across industries, yet games have often been unusually effective at turning new technologies into mass-market consumer experiences. What does the wider technology industry still have to learn from games about how people interact with digital products?

Shayan Sanyal: Games are one of the only industries where the product has to earn the user's attention every second. There is no contract, no switching cost, no inertia. If the experience is not compelling in the first ninety seconds, the player leaves and does not come back.

That survival pressure has made game developers extraordinarily sophisticated at engagement design, real-time feedback loops, progressive disclosure, and what I would call respectful onboarding -- introducing complexity without overwhelming the user. These are principles that enterprise software, financial services, healthcare, education, and many other digital products are still working to master.

Take personalisation as an example. The games industry has been doing it at scale for years -- matchmaking algorithms, dynamic difficulty adjustment, personalised in-game stores, adaptive tutorials. In the betting space, we are seeing operators adopt the same playbook: we published a guidance architecture on AWS for patron engagement that enables operators to recommend games, bets, and promotional offers tailored to individual players -- essentially the same recommendation engine logic that games studios have been refining for a decade. Other industries are only now arriving at the starting line.

What the wider technology world still underestimates is the craft. Games developers understand that the emotional response to an interaction matters as much as the functional outcome. That is a design philosophy, not a technology capability, and it is the single most transferable insight from games to any other industry.

Q: With games, streaming, social media, and entertainment platforms all vying for the same consumer, do you believe the attention economy has effectively dissolved the traditional lines between these sectors?

Shayan Sanyal: The lines have not just blurred, in many segments, they have ceased to exist. Today's consumer does not distinguish between watching a live stream, playing a game, placing a bet on a match, and chatting with friends. These activities happen simultaneously, often within the same platform or ecosystem. The mental model of discrete entertainment verticals is an industry construct, not a consumer reality.

This convergence has real implications for the betting and gaming sector. The competition for an operator's customer is no longer just another operator,  it is every entertainment experience competing for the same disposable hour. The future of entertainment is multi-dimensional, highly personalised, and deeply interactive. The platforms winning are the ones that create experiences so engaging that they become the environment in which other entertainment happens, rather than competing for a time slot within someone's evening.

You can see this convergence playing out across the industry. Major sports broadcasters are embedding live odds into their streams. Betting operators are building content studios. Streaming platforms are adding interactive overlays that let viewers engage with the action in real time. The experience of watching, engaging, and wagering is collapsing into a single surface, and the companies designing for that convergence are the ones capturing attention.

At AWS, we think a lot about this across our media, entertainment, games and sports business. The streaming era was built on cloud infrastructure. What is emerging now is what I think of as an intelligence era, where content,  interactions, and even individual frames have the potential to be personalised and understood in real time. That shift applies directly to betting and gaming. The operators who still think of their product as a standalone vertical are competing with one hand tied behind their back. The ones who recognise that they are part of a broader entertainment landscape, and design their experience accordingly,  will capture attention that the single-purpose product never could.

For betting and gaming operators specifically, this means thinking about the product as a continuous entertainment experience rather than a transactional one. In practice, the experience often begins well before any transaction takes place -- it starts when the user opens the app. Everything before and after the core transaction is either building engagement or losing it.

We are even seeing the technology layer reflect this convergence. AWS just announced dynamic multiview streaming capabilities that let viewers personalise their own multi-camera layout in a single stream, JioHotstar is already using it to deliver quad-view experiences for IPL cricket to over 100 million viewers. That is sports, entertainment, personalisation, and real-time interaction collapsing into a single product surface. The traditional boundaries between these sectors are not just dissolving,  they are being replaced by something more interesting: a single, intelligent experience layer.

Q:  AI has become available to everyone. If companies ultimately have access to broadly similar models, cloud infrastructure and data capabilities, where does sustainable technological advantage come from?

Shayan Sanyal: I would respectfully push back on the premise. Not all platforms are created equal. There are meaningful differences in the depth of AI services, the security and compliance posture, the global infrastructure footprint, and the industry-specific capabilities that sit on top. The choice of offering has a material impact on what a company can build and how fast it can move.

There is also a readiness factor that is easy to underestimate. Organisations that have already invested in modern data infrastructure, governance, and identity systems -- regardless of which cloud offering they chose -- have a significant head start when it comes to adopting AI. They are not starting from zero. The data is accessible, the security posture is established, and the operational foundations are in place to experiment and scale quickly. That maturity is a real advantage, and it is one of the reasons we see organisations with strong cloud foundations moving faster with AI than those still in the early stages of their digital transformation.

Beyond that product choice, sustainable advantage comes from three things. First, proprietary data. The models are available to anyone -- the first-party behavioural data, session-level telemetry, and player lifecycle history that make them powerful are not.

Second, speed of integration. The operators pulling ahead are not the ones with the most AI experiments running -- they are the ones shipping AI into production and iterating weekly. Speed of deployment is a significant competitive advantage that is difficult to replicate.

Third, instinct. Knowing which problems are worth solving with AI and which are better left to human judgment is itself a competitive advantage. Not every player interaction should be automated. The companies that understand the boundary between useful intelligence and unwanted interference will win on experience -- and that instinct cannot be bought off the shelf.

Q:  If technology itself is no longer the differentiator, what is? Is competitive advantage increasingly determined by what a company chooses to do with the technology?

Shayan Sanyal: It is worth being precise here. Technology has not stopped being a differentiator -- but the nature of the differentiation has shifted. It is less about whether you have access to cloud or AI, and more about the depth of your cloud offering, the breadth of your services, and how well they are integrated into the specific problems your industry faces. At AWS, we provide the building blocks -- the infrastructure, the AI and machine learning services, the published guidance architectures -- and our teams support customers as they build on those foundations. Whether it is hybrid infrastructure that can support data residency requirements, or machine learning services that operators can apply to player protection and personalisation, the capability is there for them to leverage and make their own. That is not commodity infrastructure; that is a solution designed to meet the specific demands of this industry.

That said, the question is a good one because even the best platform only delivers results if it is applied with intent. In the betting and gaming space, the differentiation increasingly comes from the product decisions operators make on top of that platform: which moments in the player journey do you personalise? How do you balance engagement with responsible play? How quickly can you launch in a new regulated market? How much of your technology stack do you own versus rent, and where does that boundary sit?

I often draw a parallel to the games industry, where I spend much of my time. In games, the tools and engines are largely commoditised. Unity and Unreal are available to everyone. Yet the creative and strategic decisions studios make with those tools produce wildly different outcomes. The same is true for operators using cloud and AI: the inputs are similar, but the outputs are determined by the quality of decision-making at the product and strategy level.

The companies I see winning -- in games, in betting, in any technology-adjacent industry -- are the ones that treat technology as a capability to be directed, not a strategy in itself.

Q: Personalisation has been discussed for years, but much of what consumers encounter is still recommendation rather than genuine individualisation. What would a truly personalised digital experience look like in your opinion?

Shayan Sanyal: Most of what is sold as personalisation today is really segmentation with a better front end. You are placed in a bucket -- high-value player, new user, sports bettor, casino enthusiast -- and the system serves you content that performed well for others in that bucket. It is useful, but it is not personal.

True personalisation would be contextual, adaptive, and continuous. It would understand not just what you did yesterday, but what you are doing right now -- your pace, your mood proxy signals, your current session behaviour. It would adjust the experience in real time: the game recommendations, the promotional timing, the tone of communication, the interface density, even the moment at which it deliberately stops trying to engage you.

On the technology side, we are closer to this than most people realise. AWS has published guidance architectures that operators can use to build real-time player analytics capabilities -- combining session-level data with ML models that update during play, so the experience adapts in the moment rather than hours later. PokerStars is a good example of the infrastructure that makes this possible. They migrated over 1,100 applications and three petabytes of data to AWS, cut their total cost of ownership by 20 percent, and increased their production change rate by 40 percent. That last number matters most for personalisation -- it means the engineering team can ship new features and iterate on player-facing experiences nearly twice as fast. Their technology leadership has said they are now building on that foundation to unlock machine learning and AI capabilities across their player data. The solution is in place; the personalisation layer is what comes next.

The missing piece is usually not technology. It is the willingness to use personalisation for the player's benefit as much as for the business's benefit. In my view, the most effective personalisation would occasionally recommend a break, suggest something different because the player seems disengaged, or simply step back. The industry is making real progress here, but there is still an opportunity to go further -- to move from personalisation that optimises for the business to personalisation that genuinely serves the player. The companies that get there first will earn a level of trust and loyalty that is very difficult to replicate.

‍

Shayan Sanyal is the Global Games Industry Business Development Leader at Amazon Web Services. He speaks at the Player Experience Academy, SBC Summit Lisbon, on Tuesday 29 September 2026. The views expressed in this interview are Shayan Sanyal's personal perspectives and do not constitute business, legal, or regulatory advice from Amazon Web Services.

Shayal Sayan and an elite line-up of world class experts will rethink where Player Experience goes next.  Secure your place at the Player Experience Academy taking place at the SBC Summit today.

‍