ARTIFICIAL intelligence deep dives in iNTERGAMINGi have only recently made their way onto our editorial calendar, such has been the rapid development of the sector in the last four years.
While the industry experts who fill our pages with their insights touch on the subject all the time – how could they not?! – taking a step back and assessing the bigger picture of AI is getting harder.
In Issue 5 last year, I wrote about Apple, Google and Samsung “firing the starting gun” on the Big Tech AI race by promoting the new capabilities of their operating systems. Anthropic’s Claude was only loosely on my radar having been through several updates since its initial rollout in March 2023. Skip forward a year and the AI sector is being dominated by models going rogue and hacking companies.
Not only has the starting gun been fired on the AI race; we are into the second lap.
The conclusion to be drawn, then, is that a year is a very short time in the AI world. To catch up on some of the latest AI developments in igaming, iNTERGAMINGi polled experts from Hub88, Mindway AI and Abelson Sports.
AI IN IGAMING
In the past year, casino aggregation platform Hub88 has rolled out a range of updates for operator clients such as My Products, Page Insights and Player Analytics. Pulling the knowledge contained within into a single interface, which also comprises Hub88’s existing aggregation suite, is HubAI, which director of product Gabriel Kolawale said enables operators to make “faster, smarter decisions.”
“Hub88 has spent years building the connectivity layer that gives operators access to content from more than 200 game providers through a single integration,” Kolawale said. “As that ecosystem has grown, so has the volume and complexity of the data flowing through it.”
Through HubAI, operators can survey player activity, transactional trends, campaign history or operational performance through “clear, actionable observations” that can support customer service, retention, risk management and commercial decision-making.
“HubAI is the intelligence layer across the Hub88 platform,” Kolawale added. “Contextual Chat answers questions about the data a user is currently looking at. The Knowledge Base integration handles product, integration and procedural questions. Player Analytics supplies the player and transactional view that HubAI interprets.
“Rather than requiring operators to export reports manually, manipulate spreadsheets or wait on analysts’ time to connect information from different systems, HubAI can interpret the data visible within the platform and surface relevant insights on demand.”
GAMALYZE SPORTS
Over at AI-led safer gambling technology provider Mindway AI, one half of the company’s flagship pair of technologies, Gamalyze, recently got an upgrade with a sports betting focus.
Gamalyze provides objective insights on gambling behaviour and potential signs of problem gambling through real behaviour, as opposed to self-reporting.
While best-known as a simulated card game, ahead of the World Cup Mindway AI launched Gamalyze Sports to turn the solution into an “interactive” football game designed to “match the immersive nature of sports betting.”
DraftKings, which onboarded Gamalyze Casino at the start of the year, was Mindway’s first operator partner for the sports betting variant, customising the solution with its own branding – including advertising hoardings.
“The response has been amazing so far,” said Mindway AI’s senior marketing manager Niamh Gallagher, who noted the pre-World Cup launch was crucial to “capture the attention of existing bettors but also the many new players that the tournament created.”
“It’s still slightly too early to share results from the tournament, but we will be compiling reports on this soon which should hold some really interesting insights,” she added.
CLAUDE IN ACTION
Sports betting data supplier Abelson Sports is not completely immersed in AI just yet but is actively using Claude to support its Football Player Markets feed.
“We recognise that there are various efficiencies that these LLMs [large language models] can introduce, specifically around log analysis,” said chief betting officer Jeeve Jeyaratnam.
“Our Football Player Markets product logs millions of decisions each day and can also, on occasion, showcase optimisations that can be found.
“The prompts and output do still require close curation and cannot be left to run roughshod over systems, so typically Abelson Sports uses Claude as a support to development, rather than a foundation. This also ensures that we remain compliant with our clients’ regulatory requirements across the global industry.”
“The Football Player Markets feed provides reliable, low latency, full live market management across goalscorer, player and team metrics,” Jeyaratnam continued. “To support this, we require immediate delivery to, and response from, clients, ensuring full monitoring of service health across the environment. Regular heartbeats provide the operator with confidence in the service availability, even during the (increasingly infrequent) times when the football event calendar is sparse.
“The service is supported both by Amazon Web Services microservices and additional bespoke services written by Abelson Sports’ developers.
“Distribution of messages internally using RabbitMQ means price, suspension and settlement calculations are sub-second, allowing operators to maximise market uptime and respond immediately to live events.”
CONTEXT WITHIN AI
Much of the optimism around the use cases for AI models is in their ability to handle complex data and multiple commands at once in the blink of an eye. Not only that, but they can return something much more coherent and useable than just a string of words or numbers.
But harnessing the power of AI is all about giving the models context – what do you want it to see, and how do you want it to interpret the data? Following what specific workflow?
Hub88’s HubAI suite puts context at the core through Contextual Chat, which can answer an array of questions about anything on screen.
Operators can ask things like, “Who is the top performer?” while a supplier query might be, “Show me the product-country distribution trend for the last four months.”
“Contextual Chat only ever sees what the user in front of it can already see,” Kolawale explained. “It inherits the same role permissions as the rest of the platform, so a support agent and a commercial lead asking the same question receive answers scoped to their own access.
“In a regulated environment the question is rarely whether an insight is useful. It is whether the person receiving it was entitled to it, and that is settled before the assistant answers.”
Mindway AI is also placing context at the heart of its new research into interactions between customers and gambling operators.
Led by Will Mace, evoke’s former director of player safety who joined Mindway in June as head of strategic programmes, the study will look at shifting the focus from solely identifying risk towards “optimising” customer interactions.
The research is being carried out using Mindway AI’s GameScanner solution, which provides detailed snapshots of gambling behaviour.
“With his background as former group player safety director at evoke, he [Mace] is already making an impact,” Gallagher said. “By combining behavioural psychology with GameScanner, it creates an evidence-based framework to intervene at the right moment with tailored messaging.
“Ultimately, this replaces blunt, blanket restrictions with precision safeguards – measuring interaction success at an individual level to improve player safety while supporting commercial sustainability.”
WORLD CUP AI REFLECTIONS
At Abelson Sports, the team is still reflecting on what was the biggest-ever World Cup, comprising 48 teams and 104 matches.
Asked about AI’s impact on the tournament, Jeyaratnam said: “Bet builders, or same-game parlays, formed a huge portion of bets taken at this year’s World Cup, and for an event with such a concentrated focus (there was very little other sporting action to compete with it), liabilities can quickly get out of hand.
“Managing correlated risk becomes very hard. AI can, and already is, tackling this issue much better than traditional risk management tools. They key is speed and accuracy, and the processing power used for AI tools is unsurpassed.
“We also saw at least one platform claim
that they traded the entire World Cup using
AI-powered capabilities. This means risk and price management offloaded to AI but also reduced betting live delays, minimised market suspension and improved bet acceptance.
“Regardless of grandiose claims, this is definitely where the market is moving. At Abelson, we still believe there needs to be a layer of human oversight, which is especially the case with our Football Player Markets props feed, which requires extensive research into football teams at varying standards – information that isn’t readily available via search.”
PARTNERSHIP FOCUS
Having focused primarily on online casino and sportsbook clients, Mindway AI has recently ventured into the land-based gambling sector with Crown Resorts – the company’s first deal in Australia – and daily fantasy sports with PrizePicks.
These collaborations have “significantly broadened our technical capabilities by exposing our AI to vastly different player behaviours and data environments,” Gallagher said.
She added that the deals mean Mindway can test and “sharpen” its machine learning capabilities against “entirely new” risk indicators and player touchpoints, which, it is hoped, will benefit its technology’s use cases across verticals.
“Ultimately, this cross-channel experience transforms Mindway AI from a specialised online detection tool into a highly adaptable, full-spectrum safer gambling ecosystem capable of delivering precise, real-time player protection across any regulatory or operational context,” Gallagher said.
Abelson Sports’ key partners include AWS and its support partner DoIT International.
“We utilise their platform tools to benefit from existing expertise: SQS for message routing, ECS for application containerisation and resilience, and Route 53 for DNS routing,” Jeyaratnam said. “Historically, Abelson Sports used Rackspace cloud provisioning extensively but chose to move to other service providers following a perceived reduction in their service. This led to some of our services being moved to Digital Ocean, where the complexity of AWS provisioning was not required.”
HUMAN INTERVENTION
It is easy to get carried away with the capabilities of AI while forgetting the quite crucial role humans will have to play in plotting the points of its evolution carefully.
“While HubAI can significantly improve speed, accessibility and efficiency, some situations still require human expertise,” Kolawale said. “Complex technical issues, account-specific investigations, commercial discussions, regulatory matters and other operational scenarios involve factors that go beyond the limits of what an automated assistant should handle independently.
“In these cases, our support team remains available to work directly with partners and provide personalised assistance through service desk tickets or direct escalations from their account managers.”
For Gallagher, she said Mindway AI “firmly believes” that AI should “empower, not replace,” human expertise. This means the company’s Customer Success team and independent expert panel of clinical psychologists will remain on hand to intervene when required.
“Most recently, Australian psychologist Susan McNulty joined the panel, bringing over 20 years of experience in harm minimisation and the strategic management of problem gambling within large-scale casino operations,” Gallagher said. “This combination of frontline customer success leadership and world-class clinical insight ensures operators always have direct access to human expertise whenever nuanced guidance is required.”
MORE AI DEVELOPMENTS
Among the other recent AI developments in igaming is research from live casino game provider BetGames, which covertly replaced human dealers with AI replicas in one of its gameshows for several days.
The company said fewer than 30 per cent of players noticed the hosts were AI-generated and that there were “no significant deviations in key performance metrics.”
However, BetGames said there is not yet a strong enough business case to roll out its technology at scale after noting a lack of both positive and negative movement from the avatars.
The supplier said there is currently “no substantial financial upside” of AI presenters and that the practice remains “resource-intensive.”
The research, published in March this year, came before FanDuel co-founder Nigel Eccles spun out the AI live dealer technology trialled on his BetHog crypto casino. Sentient Studios’ tech is now available on a B2B basis for other operators.
Meanwhile, a survey by Censuswide on
behalf of AI product builder Studio Graphene found a perceived lack of return on investment from AI – and confusion about whether AI has worked.
Censuswide polled 500 decision-makers within UK businesses and found that, while 78 per cent of the companies are using AI, less than a third (31 per cent) have seen positive ROI.
Almost a fifth (18 per cent) said their AI projects have not delivered the benefits they expected. Less than half (41 per cent) of AI users said they have a clear idea of what success looks like when using the technology.
“There has been a rush to adopt AI amidst huge hype and a proliferation of new tools – this is certainly true of private equity-backed mid-sized companies looking to AI for automation, scalability and competitive edge,” said Ritam Gandhi, director and founder of Studio Graphene. “The problem, however, comes when AI is deployed without first defining where it sits within workflow, the decisions it’ll inform, the processes it’ll support and the criteria for measuring success. Often, teams haven’t agreed whether AI is meant to save time, improve decision quality, reduce risk, support growth or all of the above.”
It is clear then, that while AI has the potential to transform all of our lives, there is plenty of potential to realise – and there are a similar number of concerns to address.
