By AI Trends Staff
Operators of casinos and online games are incorporating AI in efforts ranging from maximizing profits to helping problem gamblers.
The gaming industry is technically savvy, having integrated automation into its operations to gain efficiencies and offer conveniences to customers. Now AI is being applied to casinos and the gambling industry, in-person and online, enabling more advances such as allowing multiple users to play the same game at the same time from different locations.
Other advantages include ability to track compliance with online gambling regulations, collection of data on gambling preferences to enable predictions and deliver customized service to customers, according to a recent account in LA Progressive.
It might be difficult for operators to enhance the customer experience without the use of AI in the future, suggested a speaker at SBC Summit Barcelona – Digital, the Global Betting & Gaming Show, usually held in Barcelona but held online recently.
“For sure what we are focusing on is to increase our customer experience, the winning experience, and that leads to revenue and revenue on the lower end. I’m not seeing operators managing their operations without the help of AI in the near future. Talking about revenues, I think it’s too much over the next three years with everything on the internet,” stated Américo Loureiro, director of the casino firm Solverde, a casino and hotel firm, in an account in CasinoBeats.
“Our plan is for the next three years to increase the AI on our operations and get benefits from this. We know that this is the very beginning, and we want to be on top of this because the ones that agree more with AI and manage AI will be the most successful operators.” he stated.
Startup Rootz was formed in 2018 by internet gaming (igaming) professionals intent on building an online gaming platform. “It was most definitely a strategic decision to place AI in the center of thought when we started designing the platform,” stated Edvinas Subacius, chief data officer of Rootz, speaking at the SBC event. “We know that we can increase player turnover or spending up to five per cent by doing recommendation engines. At the same time, we know that we increase their lifetime value between ten per cent to 20% so four times more than actual spending by applying AI to manage our bonus cost and promotional values.”
He suggested trying to weight the benefits of AI by a headcount measure of the number of operations the casino can get done per headcount, or how much additional revenue the headcount generates. Subacius stated that his company has 70 employees, “and we are running operations equal to other companies that have 300 to 500 employees. So the sooner you start with AI, and the closer it is to the heart of your mentality and your platform, it can create the efficiency of 100% or 1000% scale. It’s not 5% anymore.”
The gaming industry has been collecting data from customers and market for years, and now the industry is positioned to use AI and machine learning to advance business goals, suggested Steven Paton, business solutions advisor for BMIT Technologies, a multi-site data center provider in Malta. “The next step is definitely the transitional phase of utilizing big data and machine learning to push that further over the next five years,” Paton stated at the SBC event.
Norsk Tipping Constructively Works with Problem Gamblers
The risk that AI will more effectively target problem gamblers is recognized by Norsk Tipping, Norway’s state-owned gambling company, which is working with a software provider on behavioral analysis to help identify problem gamblers.
The 2.2 million customers of Norsk Tipping have passed identity checks and have gaming accounts, which track game-playing frequency and winnings received, enabling the company to set gaming limits if necessary. “This wealthy supply of data provides unique possibilities to exploit AI to prevent problem gambling. Norsk Tipping’s mandate is clear: the company shall act to prevent the negative consequences of gambling,” stated Tanja Sveen, advisor responsible for gambling for the company, on the blog of Norsk Tipping.
Machine learning can help expose risky gaming patterns and personalize preventive measures to reduce risk gambling, she suggested.
Norsk has been working with a gaming behavioral analysis tool from Playscan, a company founded to address issues around problem gambling. The analysis tool aims to expose risk-filled gaming patterns, provide the customer with feedback explaining the reasoning behind the risk assessment, and suggesting preventive measures where appropriate.
The first Playscan model was based on a number of self-assessments completed by customers. Those were used to develop the second-generation model, by comparing customer data from the first period to their responses to a self-assessment questionnaire for the second period, in order to train the model. “More than 60,000 completed self-assessments were used to develop the second generation of the Playscan model,” Sveen wrote.
Results so far have been positive. “The new model is clearly better than the previous one. It has a higher level of accuracy and the customer feedback shows a greater degree of agreement with the risk assessment,” she stated.
Proactively Calling Problem Gamblers
Norsk is using AI in an effort to proactively call—on the phone—customers with problem gambling habits. During the call, the customer is given facts on their gambling spending and the need for change is discussed. This often results in a reduction of the customer’s gaming limit.
In order to identify which customers would most benefit from a call, the company set up a study on machine learning with the BI Norwegian Business School during the spring of 2018. The researchers used a sample dataset of 1,400 customers who had received a proactive call, a random selection from the 10,000 customers who had lost the most money through gaming in the last year. That gave the team a representative data set useful for creating a model. Customer’s theoretical loss 12 weeks before the call was compared with their spending 12 weeks after the call, to see if spending had decreased or increased.
The standard procedure for machine learning was employed; the model development and the data evaluation was carried out using the automated machine learning tool called DataRobot.
“The evaluation showed that we managed to develop several models with an ability to provide rather highly accurate predictions,” Sveen wrote. “To a great extent the models were able to identify the customers who made use of the proactive calls.”