- Upcoming event:
- Moscow, Russia
- Add to Calendar
Andrew Pearson shows how machine learning helps combat gambling addiction
Advertising that follows clients from a website to a website and adapts to their sex, age, and interests is a marketing trend over recent years. According to Andrew Pearson, personalization may be improved much more. Artificial intelligence is an answer.
In his interview, President at Intelligencia Limited, a company developing innovative technical solutions for casinos and betting operators, told about AI technologies: types of machine learning algorithms, how to apply it in gambling. He also commented on the related discussion at Russian Gaming Week held in Moscow.
Interviewer: RGW Moscow (RGW)
Speaker: Andrew Pearson (A.P.)
RGW: Your career included the cooperation with major casinos. Could you please give examples of how AI
A.P.: AI and ML can be broken into three types – supervised learning, unsupervised learning and reinforcement learning. Unsupervised learning breaks down into clustering and dimensionality reduction, with the former being more important to casinos and sports book operators. Clustering allows for customer segmentation, which, with AI and ML, can be done almost in real-time as models on feeds from multiple data sources can be created on-the-fly and one can almost have a ‘living segment’ that instantly evolves as customer interactions and customer behavior on social media are taken into account. This can lead to better target marketing as well as help build better recommendation systems.
Unsupervised learning also allows a casino to look at its entire customer database and find specific personas and put people into highly specific buckets so that not just the marketing message but also the communication methods and timing can be set so that the marketing message is highly personalized and set to go out at the most appropriate time.
AI is also going to help with online advertising, taking it from a programmatic method to a cognitive one. In the past, it always appear as though advertisements were following you around as you went from one webpage to the next. What happens with cognitive advertising is the system is intelligent enough to understand the reason why you left a site, how you left, whether you bought or not; basically, it will make an informed decision on whether ads should be served up or not. AI will bring a whole new level of intelligence to advertising that will be noticed by the person who sees the advertisement.
Finally, unsupervised learning can help with fraud prevention.
On the supervised learning front, you can build problem gambling models. Looking at customers who have opted-out because of problem with their gambling habits, AI can spot patterns that might go undetected to a human in terms of their gambling behavior or their demographics and potential psychological history. Supervised learning can also help with customer retention, identifying key patterns in purchasing, as well as marketing forecasting and figuring out customer worth.
With reinforcement learning, AI and ML can be used to capture real-time cheating, as well as help with skills acquisition and learning tasks.
RGW: What competitive advantages do gambling companies obtain from AI applications?
A.P.: ‘Personalization’ is a word being bandied about in marketing circles these days and AI can help enormously with personalization marketing. The entire marketing process was radically changed with the introduction of marketing automation over the past few decades, but now we’re moving into AI marketing that is going to increase . The AI-powered marketing way is to let the system decide what is the most optimized way to reach a customer so that the marketing message gets through. We’re went from sending out one-size-fits-all marketing, filled with one set of images for everyone to a more personalized message, i.e., an email filled with images that reflect the marketed-person’s desires and tastes, to an AI-powered marketing system that keeps up with that individual’s behaviors and social media posts so that it can market messages at the time when the marketed individual is most receptive to them. AI is given the goal of ‘When is the most likely time that an individual will open a marketing offer?’ and it gains as much information as possible about an individual and then sends out the offer when it considers it most appropriate.
RGW: Tell us in detail about intelligent solutions used for data storage. What data do casinos collect and process? How will AI help to optimize the process?
A.P.: Currently, casinos and IRs are collecting an enormous amount of data on their customers, an almost overwhelming amount and without AI and ML they wouldn’t be able to keep up with it all. Real-time streaming will become an important part of customer interactions and without AI and ML, it would be impossible to capture a message, utilize it o create a marketing piece and then fire it off in time for it to be useful. AI will help less in the storage of the data and more in the use of the data. AI can recognize where the data is, keep it in places where it is most useful and then track it as it evolves. AI has the ability to learn on its own so things like fraud detection can be initiated through AI and it will continue to grow and evolve, capturing things that would never be spotted in the normal way.
RGW: Which of the gambling areas (land-based casinos, virtual casinos, sports betting) are the most promising for AI integration? Why do you think so?
A.P.: AI can help all of these areas, but I’d say it might be best for sports betting because, in general, analytics is a more useful for sports books than for land-based casinos. Sports book are dealing with very thin margins and a lot of competitions, therefore they need to bring personalization and marketing to a whole other level. AI can increase customer understanding to such a detailed level that.
RGW: Are there any companies specializing in the development of AI systems for gambling?
A.P.: Ever one of my software partners are now specializing in AI and ML. These two technologies are seem as growth areas, even countries are setting up initiatives to increase funding and incubating companies as they see it as one of the next big technological areas. My partners TIBCO, Alicloud, SAS, and SAP all consider as well as BI vendors.
RGW: Thanks for your presentation at RGW (Moscow). What are your impressions of the event? What do you think about the audience’s activity and involvement during your presentation?
A.P.: Highly positive event, very professional. I particularly like the social media marketing that was done. The language was obviously a barrier, but the audience seemed to understand and be open to my message, even laughing at the right places. Few questions at the end, but I assume that’s because of the clarity of my talk.