AI is reshaping how integrated resorts can identify and benefit not only premium customers but also grind mass, breaking the bounds of traditional tiered loyalty programs.
“Data soup.” While not the shiniest term, the description by AI expert Mana Azizsoltani encapsulates what integrated resorts are currently placing on boil, gathering an impressive amount of data points that run the gambit from side bets on smart tables to almond milk in morning coffee.
While regulations have yet to catch up, the possibilities around the intersection of AI data interpretation, facilitated by smart tables and facial recognition, can allow for the tracking of a player across an entire property: every wager, every purchase, every exuberance or discontent.
Starting at the gaming table, what could this mean?
“Looking at something like customer acquisition, you can use AI to detect players that are high value on the floor that might not be carded, might not be engaged, and go find them where they are and sign them up, talk to them, offer them a room,” notes Azizsoltani, Head of Analytics at Differential Labs.
Aaron Raj, Chief Technology Officer at Angel Australasia, says this initial player information is essential, with smart table technology and facial recognition providing valuable additional insights.
“The system tracks every human element at the same level, same depth, whether he’s carded or uncarded, registered or unregistered,” he explains. “So, the patron doesn’t even know that he’s being tracked at that level. In this form, the way you do your loyalty completely changes, and how you address it is at a different level.”
Speaking on a panel at the IAG Academy Summit in Manila, Azizsoltani and Raj were joined by Light & Wonder Systems Sales Manager, Asia, Herris Kocibelli, who clarified that loyalty programs are about to undergo a massive shift.
“It’s not about a tier anymore, it’s about a ranking,” he offers, pointing to the ability today to capture anonymous play. Every time a player places a bet – be it on a smart table or on a slot – the information is captured and the individual assigned a number. Their information can then be rated accordingly, even as an uncarded patron.
“You’ll have a specific number with which the operator will be able to tell how valuable this particular person is to me,” notes Kocibelli.
This upends the traditional loyalty program format, eliminating the road bumps of sign-ups and immediately giving value to players without them having to show initiative.
Speaking on a separate panel at the IAG Academy Summit, Joe Mayer, Systems Sales & Operations Manager, APAC for Konami, noted, “If you can identify and reward every single player that comes through your door and increase your VIP journey and customer satisfaction, it pays for itself,” highlighting the pay-off on investment in both facial recognition and AI.
But IRs have yet to fully catch up, notes Syntrii Founder and CEO Laurent Fresnel. The expert called out operators for still taking a siloed approach, noting that “the customers are more than the people that you know” and urging them to apply a unified approach to management in order to track all of their customers’ data points and identify their behavior across-property, and even online if the operator is omnichannel.
Bryan Jenkins, Director of Angel Manufacturing Singapore, dialed in on the fact that online has always been data-driven, with every interaction recorded.
“Online gaming has always had a huge advantage in that the information is digital and is immediately available,” he said. “Every transaction, every touchpoint has generated some sort of digital record that can be retained, where traditionally on table gaming that wasn’t the case. It’s only since the advent of smart tables that you are really able to digitize that information.”
Both online play and slot machines benefit from the confidence that operators have in the data being provided, while table games have previously required a much higher level of oversight, with table supervisors, pit bosses and constant checks on wagers placed, amount paid out and whether a player is actually allowed to wager or not.
Smart tables are eliminating that necessity, but the confidence isn’t quite there yet.
Kocibelli notes that “We can’t just rely on AI to give us what we want and then blindly trust it. We still need to keep an eye on things.”
Referencing Angel’s hybrid smart table system, which utilizes a combination of RFID and AI capture, Raj further stated, “AI is never 100% accurate. How do you manage that uncertainty? It’s not all about data from a smart table perspective. The data is there, it’s granular, and then you are supporting it with snapshots, pictures for every hand from the time the card was drawn until it’s discarded. The player, when he’s placing his bet, everything is snapshotted. So, there’s a whole bunch of images for every hand. And then on top of it, you also have video from the time the bets were placed by the player until he walks away. There’s a complete video part of it.”
Raj notes that the first element is the “data layer that was created,” which is further enhanced by the “evidence-based layer” – itself interpreted by AI.
“It’s not purely relying on AI because what you feed in is what you get,” he says.
“The important differentiating factor is that just by bringing in technology, a table does not become smart.
“AI needs to understand the game, convert that into data and make the data available for the operator to make decisions. In that way, the operator is able to get the complete benefit of it. The player is rewarded in terms of the exact bet being placed, that’s converted into points and comps, and in that way it all comes together. It closes the cycle. AI is there in the background, invisible to the player, unobstructed to the operator and indispensable for the operation.”
But that indispensability comes with indispensable staff.
Azizsoltani laments that the “brilliant ideas” and “roadmap” that operators have for AI implementation is hamstrung, because operators “can’t attract the talent.” The young expert highlights how Silicon Valley and tech are more magnetic options, even though “everybody wants to do AI.”
Even when operators have the AI experts on board, he challenges that “the data is not really all in one place.” Azizsoltani notes that operators tend to first look for AI as a solution, before identifying the problem.
Says Raj, “All these different teams, be it cage or loyalty, marketing, surveillance, analytics, IT; all of them look at a disparate set of systems. But if you look in the background, it’s the same set or action which has happened, but they’re looking at it from a different set of systems.”
For Kocibelli, the biggest obstacle to greater AI adoption with the industry is that regulations are an ongoing hindrance, even though his panel praised how Macau is moving forward quickly and the Philippines has been very swift to adopt. The Light & Wonder executive highlighted that “there are regulatory restrictions that don’t allow you to do certain things,” positing a future where a patron can be identified by facial recognition at a marketing or club booth, staff can approach, greet the player by name, know their history and how to offer them a refreshing and engaging gaming experience.
But that gaming experience can be taken advantage of.
Azizsoltani notes that the recent proliferation of baccarat side bets, enabled by smart tables, are being identified by advantage players as an opportunity, with AI also helping them maximize their advantage.
In a recent report he helped pen, Differential Labs found that teams of exponents are now using AI to identify when these side bets offer a mathematical edge over the house, often calculating many side bets at the same time, with leakage across Asia estimated at between US$500 million and US$700 million annually for bricks-and-mortar and online casinos combined.
“The change in the odds is just ginormous,” Azizsoltani highlights, pointing to a type of “AI arms race” between operators and players, as one side tries to game the system and the other tries to keep it fair.
Azizsoltani says that with generative AI, “the operator has more to gain,” at least from a table games perspective, taking a positive spin on the battle as a “democratization of analytics.”
But Kocibelli raises a particularly terrifying possibility for the future: self-teaching AI programs that go far beyond what regulators could hope to keep up with. His suggestion: “instead of regulating technology itself, I think we should probably regulate more on the outcomes of what the operator is actually trying to achieve.”
No matter what, AI is here to stay, but how it’s implemented will continue to be a race to the top, with massive advantages, many concerns and a lot of excitement.
