AI: Disrupting the Retail Casino

Artificial intelligence has become part of our daily lives, at home, on the job, online and in business. The technology, trained to mimic human thinking and decision-making, has been presaged for decades, in sci-fi movies and Twilight Zone episodes.

But it wasn’t until November 2022, when OpenAI introduced ChatGPT, that the future became present tense. From launch, the generative chatbot had a seismic impact, reaching 1 million users in days, and sparking an AI arms race, pitting Microsoft and Google against OpenAI and lesser competitors.

At the very least, AI offers casinos better ways to harness and interpret consumer data, virtually in real time, for more responsive and personalized customer service. Theoretically, that means happier customers who stay longer, play more, and return like boomerangs.

But data analytics are the tip of a massive iceberg, with the potential to upend decades of business-as-usual. GenAI “is changing every single aspect of technology in the casino space,” says Andrew Cardno, co-founder and CEO of Quick Custom Intelligence, a global leader in business intelligence.

“Start with game design—the artwork, the design, the math models, the planning. Game designers today are building games at a rate that’s 10 to 100 times more productive.”

Likewise, AI is expected to transform software tooling, customer security, responsible gaming, marketing, compliance, recruitment, staffing—the works.

Along with complex jobs, turning mountains of data into actionable insights, it can perform “simple but mundane tasks—automating the slot audit, as an example,” says Jon Wolfe, president, global systems and services for Light & Wonder. “You may go, ‘Well, that seems pretty easy.’ But it’s important and it’s impactful. Some of our operators spend a lot of time and manpower doing these audits.” Automation “frees them up so they can use that capacity to drive more important things.

“AI is doing these tasks better, more intuitively, more reliably, so we can focus on complex decisions that are more critical to the long-term viability of the casino.”

From Uncertainty to Excitement

Wolfe was an early evangelist for AI as a tool in casino operations.

“At first, it was a lonely conversation—just us promoting the technology, and a lot of uncertainty from operators who weren’t sure what this meant,” he says. “Trust had to be built up around the technology and the data.”

Historically, gaming has been slower to adopt new technologies, in part due to the plodding pace of regulatory approvals. But in this case, acceptance on the operator side came quickly.

“I’ve got to give credit to the industry—it really went to school on AI,” says Wolfe. “Now operators are showing up full-force, not just with questions about AI, but their own ideas about the role it can play on casino floors, in strategic determinations about players and operations. They’re coming to us going, ‘I need predictive maintenance on the management of my bill validators based on volume.’ And that’s great. This is one case where the industry leaned in, in a really meaningful way.”

Across the business spectrum, initial doubt about AI seemingly has been replaced by a new wave of “techno-optimism.” According to Scientific American, 96.9 percent of businesses now view AI as mostly beneficial, “a once-in-a-generation transformational moment, akin to the founding of the internet in the 1990s.”

However, the World Economic Forum reports that just 16 percent of companies consider themselves prepared for AI-driven reinvention. And 74 percent acknowledge “critical barriers to scaling AI solutions.”

Player Development vs. Player Safety

Trained on player data, and based on individual behavioral patterns, AI creates granular consumer profiles, from wallet and credit status to gender and age. It can recognize their distance from a property, and when they arrive. This tracking “factors into what a customer will do, how long they’ll play, and how they’ll react” during a visit, says Wolfe.

Light & Wonder “built almost 2,000 iterations of that model because of the AI tool set, and did it very, very quickly: adjust, iterate, adjust, iterate. We could predict on an individual basis when people were reaching the end of their gaming session, and incentivize them in real time to extend their play.

“So AI really does let you get into some very advanced mathematical items that used to take months. We could do a model in minutes, evaluate its viability, take a step back, recalibrate it, and relaunch the question.”

If that seems almost Pavlovian—ring a bell, and wait for the experimental subject to salivate—Wolfe says AI also can be built to detect when gambling gets out of hand.

“Responsible gaming is where we built some of our most viable models—it’s almost the exact balance of marketing. We can look at individual habits and responsible gaming models to see when somebody’s engaging in risky behaviors—on an individual basis rather than setting some very broad rules.

“The role we can play in the operational system is anywhere from coach to cop. We can send out an alert, have a host intervene or actually shut the game down. That’s a pretty exceptional opportunity.”

AI can also enhance surveillance to detect potential threats or cheating, for additional security.

“As we’re driving players to better levels of loyalty and participation,” says Wolfe, “we can use that same technology effectively to protect the player from negative opportunities.”

Still Experimental

But it’s not a quick remedy or one-size solution. As Alvin Toffler noted in his landmark Future Shock, published way back in 1970, “Every statement about (technological advances) ought, by rights, to be accompanied by a string of qualifiers—ifs, ands, buts, and on the other hands.”

“You need to know where you’re trying to go, what problem you’re trying to solve,” says Wolfe. “AI plays an invaluable role in getting there quickly. But it’s not the Magic 8 Ball.”

It’s also only as good as its training. Eric So, professor of global economics and management at MIT’s Sloan School of Management, has written, “We’re in an experimental phase, where it’s important to think about the problems you’re attempting to solve and the sensible business applications of AI.”

And there are pitfalls. These humanoid technologies may actually echo cultural biases. “There’s quite a bit of research showing that certain types of (large language models) can present favoritism toward certain ethnic groups,” wrote So. “This is a really good example of the importance of careful monitoring of LLM outputs and understanding both their capabilities and their limitations.”

Of course, there are concerns that AI will replace human beings in the workplace. According to a 2023 Goldman Sachs report, AI could replace 25 percent of work tasks in the U.S. and Europe and slash wages in many professions. Many once-secure jobs are no longer safe.

But the technology may also create whole new categories of employment, do hard-to-fill jobs to address the post-pandemic labor shortage, and reduce the cost of services.

“A lot of the cognitive skills we currently do, we’ll push to AI,” says Cardno. “But human creativity and the human mind will expand to do the things that we’re really good at. There’ll be a whole new set of roles and new tasks that will just emerge.”

The Next Industrial Revolution

As AI evolves, it will shift from analyzing historical data to future-casting, observes Wolfe.

“We’re working on using AI to formulate what your marketing should be for the month to generate the optimum amount of revenue: what are the promotions, who should be invited, hotel occupancy. What’s coming down the road? What’s my revenue? Those are really complex models, more predictive in nature.”

Robots may be next—not just robotic cocktail mixers or vacuum cleaners, but walking, talking, human-like robots capable of service tasks once performed only by people. According to Oxford Economics, breakthroughs in reinforcement learning have already equipped physical robots to perform complex physical tasks like hanging T-shirts on hangers and kneading pizza dough.

“The change is transformative, gargantuan, unstoppable,” says Cardno. “It’s happening. As a Star Wars fan, I’m wondering who’s going to be the person who builds the real C3PO.”

While these visions may sound apocalyptic to some, Cardno says they’re “more 100 percent transformative. If AI isn’t part of your current work plan, you won’t be able to compete in three years. You may as well be a fax company.”