To 2020 and beyond: the future of AI in igaming

With Digitain’s Director of International Development, Simon Westbury

Our interpretation of artificial intelligence, its progressive role in the world, or its actual relevance to us in any given moment very much depends on our domain of discourse. We’ve all seen the dystopian movies about the rise of the robots. Elon Musk is talking about self-driving cars soon rendering human-driven vehicles illegal, at least on public roads. One AI “teacher” even apparently improved student marks by 35% compared to its human equivalent – wish I had one when I was at school! Meanwhile, back in our own realm of igaming, just take a stroll around any conference floor and you’ll see the majority of stands promising to automate you to the moon and back – and walk your dog while they’re at it. Surely they can’t all be right, at least right now.

So, let’s dispense with fantastical future, and deal in more sober realities. Where are we presently at when it comes to AI? Well, having chaired a panel on the very subject at the recent Eastern European Gaming Summit (EEGS), I’m now in a position to corral and share the thoughts of some top industry experts on artificial intelligence and its current implementations across our industry.

Automated, algorithmic trading is perhaps the most obvious use case for the industry, particularly for in-play betting which has taken over as the main driver for turnover in the UK (up to 70% in some reported cases). However, dig a little deeper or speak to any senior sports trader, and they’re still loading up and Excel spreadsheet, enabling macros on formulas, and overseeing the games themselves.

Quantitative models, running off data channeled from global odds markets, can accurately adjust prices and manage risk as things unfold. However, that’s it for now. By way of analogy, take chess. The algorithms may now wipe the board with the Grandmasters, but the best human-computer teams still consistently win against the robots. Algorithms can process a myriad of moves, but a more detached risk-management system with man-and-the-machine oversight still has its place. As with every intelligent endeavour, though, AI systems will surpass us one day. So, the costs of running sportsbooks are forever diminishing.

For my part, I’m especially interested in CRM as the next big potential differentiator for our operator partners over the coming year, especially in Europe’s more mature market post-GDPR. To which end, I predict that 2020 will be the year of the player journey and player protection. Therefore, it’s now up to operators and suppliers to process the myriad data points available nowadays, whilst also securing the best way to stay fully compliant within the tightening grip of worldwide regulation. Which means that both operators and suppliers must start asking the right questions – even basic questions such as: what do I want AI to achieve? I worry some industry peers sometimes expect the solution without taking the trouble to pose the correct question.

Expect the trend towards the personalisation of marketing to pick up its pace. It’s all about leveraging the data at your disposal, converting it into actionable insights that can boost your bottom line. Only optimal behavioural data insights and personalising your communication will reliably unlock the door to “dwell” time, not to mention gaining ground on the holy grail goal of knowing your customer from the login – their passions, their proclivities, the offers and bonus that drive engagement, a customised user interface promoting the markets they want. Tags which flag behavioural traits (in-play biases, preferred sports, even a customer’s favourite club or player) also offer a personalised and customised experience, configuring all aspects of the customer journey in a responsible, trackable way.

Old-school marketers may baulk at some of our revolutionary code-reliant methods. But at the end of the day, this is simply a question of data processing by a bigger-brained beast. After all, efficient information processing represents the backbone of any intelligent system, and no CRM team can handle these modern-day challenges manually. Basically, AI is a tool to hone and gear up what you can already do manually in order that a more efficient process can ensue. That’s true whether you’re measuring margin, safeguarding your customers from high-frequency deposits, or installing an optimised marketing campaign to better speak to the customer and elevates their experience.

There’s also a danger that regulation is becoming so politicised that new technologies, led by AI, are not being fully adopted in a way that best benefits the customer. That’s because regulators simply don’t understand AI since their respective CVs do not include any tech or gaming know-how. So, we need to promote effective use cases in order to better articulate the AI proposition and its inherent advantages to the relevant authorities.

Ultimately, my view is AI needs to be fully understood in the specific areas where we seek to employ it. We need to focus on the player in terms of entertainment and protection. We need to cooperate with the regulator in terms and language they quickly comprehend. One challenge which the entertainment AI disruptors of Netflix and Amazon have is that they cannot sell cross-platform. In our own industry, from an operator POV, we can. Accordingly, we’ve since seen a lot of stuff around request-a-bet and build-a-bet formats, so we’re getting some sort of worthwhile personalisation. Nevertheless, we can perform far better as we embrace machine-learning techniques. In fact, this represents one of Digitain’s central focuses – understanding the wealth of data we have and using the algorithms to produce a gaming equivalent of the Netflix experience, Betflix so to speak.

I concluded my seminar in Sofia by asking the panel to assign a percentage as to where we currently are when it comes to exploiting AI in gaming. Their answers varied between five and 20%. And while such predictions invariably aren’t worth the paper they’re written on, considering tech’s take-off potential (please pester a professor to explain the doubling power of Moore’s law!), we all agreed that we have barely begun to scratch the surface here.

Whatever the immediate future, it’s an exciting time at present if you’re operating in one of the most progressive and tech-friendly industries around. We’ve now put our powerhouse sportsbook platform in sync with these advances in automated processing. Those operators marooned on legacy technology, however, now have a new raft of agile bookmakers nipping at their heels. At the same time, others who own their own tech are capable of pulling away from the field. If I was stuck on a legacy platform, I’d be very concerned right now.

Author: Simon Westbury