On the data hunt: how AI became the new normal

Artificial Intelligence, or AI as it is more commonly known, still holds the air of science fiction about it. The whiff of robotic masters, learning and growing and becoming omnipotent. In reality, the fact is that AI has integrated itself into daily life seamlessly, without the Terminator-style entrance.

From apps such as Uber to copy-checking software like Grammarly, AI functionality is part and parcel of the world we live in today. Similarly, across the gambling industry, AI now acts as a data-driven glue that holds various sections of the industry in place.

From odds creation to customer service, onboarding to screening, AI is able to be deployed across every step of the customer journey, with a key focus on speed and efficiency.

Data is king

For the likes of data giant Stats Perform, AI is at the heart of its approach. Patrick Lucey, VP of AI, tells EGR Technology that the use of AI is now a base requirement in the modern world.

Lucey says: “AI is the new normal for all businesses across the globe and permeates across every sector. Every business is now a data business, with every element tracked by data.” The crux here is that without the data AI will ultimately fail. As how a great striker relies on their teammates to fashion chances, AI is dependent on its in-fed data to perform adequately.

At Stats Perform, a team of around 50 people, mostly with PhDs, work in the AI team with a trident focus. Predictive modelling looks at team and player performance predictions, computer vision entails deriving data from video and natural language generation uses AI to creates stories and insights based on the data.

Lucey notes: “To do any type of meaningful AI, in any industry, you first need to have a large amount of data. The more data you have, the better the AI technology will be – simple as that.”

For operators, this rich hoard of data funnelled into AI systems infinitely improves its offering. Thomas Alomes, head of North America for the Sports Tech World Series, tells EGR Technology that implementing this technology can improve the integrity of an operator’s product. “AI enables the automated collection and analysis of sports data, thereby creating operating efficiencies while bolstering the integrity of data,” he says.

Odds-on

One of the key areas AI can support bookmakers is in its odds creation and modelling, across both macro and micro scales. In terms of odds creation, this aspect of AI is known as machine learning, whereby the AI learns directly from the data that is has access to, instead of being dependent on human input.

By accessing large swathes of data, AI can delve into smaller, player-specific markets, which Lucey argues will only improve the customer experience. “Sports data reconstructs the story of a match – the deeper the data you have, the better the reconstruction and the better the analysis of specific facets of team play or a specific player’s performance,” he says.

“That means AI can power betting experiences that can be confidently extended into many more player prop market types, for more players, so bettors aren’t limited to only a handful of players and props,” he continues.

Some might argue that a betting market on every action a player makes on the field could lead to oversaturation, but the sheer number of markets available to consumers today would suggest there is a tangible taste for what AI can provide.

Elsewhere, the almost instantaneous changes in sportsbook front-ends are due to the capabilities of AI. Lucey details how the speed and efficiency in updates from a user experience standpoint would be nigh-on impossible with a human alternative.

“Instead of waiting for a human operator to consider the implications and adjust the odds for every market due to a pre-game or in-play change in a match, AI models can update immediately and are self-consistent as they come from the same kernel of knowledge. This is beneficial because it limits the time the market would be suspended, again improving bettor experiences,” he says.

Manual is mandatory

In terms of responsible gambling, advanced AI can be used to upgrade on standard KYC practices in the industry. For William Hill, which onboarded financial crime screening service Accuity earlier this year, the use of AI insights aims to shore up its already robust screening process.

Accuity, whose services have been used by the banking industry, screens customers in real time and identifies high-risk individuals based on a set algorithm, with the firm stating it can reduce false positive matches to below 1%, compared to between 10% and 15% as a gambling industry average.

At the time of release, Hills’ AML director, Steven Armstrong, said that the partnership would improve accuracy, “without adding to our manual workload”. Armstrong added: “AI technology will never replace AML decision making, but it can certainly improve it. By minimising false positives, we will be able to reallocate resources, rather than managing an excess of false-positive flags.”

Lucey, despite labelling AI as having “superhuman capability”, agrees with Armstrong in that the human element to processes needs to remain in place, and AI should act as a supplementary benefit and not an eventual replacement.

“We do not have any fears that AI will take jobs away. We see AI technology as enabling human experts to do their jobs more effectively and confidently and create better products and experiences which can service more customers, faster and more sustainably,” he remarks.

Alomes, however, states: “Traditional manual methods of sports data collection and analysis are labour intensive and open to human error. They require people sitting in the stadium watching the game, manually reporting what they’re seeing.”

What now?

If the future of AI does not lie in replacing those who created it, and instead acts a data-fed tool to minimise workload and improve efficiency, what are the next steps for the technology? For Lucey, as ever, it goes back to the data.

“AI technology is only as good as the input data. Even though we have an enormous amount of sports data, we still need more, to understand and predict the action and impact of every player on the field. To that end, we want to collect even more data and capture it at a granularity that a human could not do manually,” he says.

There could be an argument that data is essentially finite, you can’t summon organic data out of thin air, but Lucey details how to expand collection, becoming a trawler net to pick up even the smallest of details.

“Our vision is to travel back in time and collect tracking data from every sports video that has ever existed, which is giving us even richer data for our models. We have started with capturing tracking data for every NCAA college basketball game ever played, as well as soccer, but we are scratching the surface of what this technology can do.”

Scratching the surface is both terrifying and exciting at the same time. AI’s capabilities, not just in the gambling industry, but every aspect of human interaction, is a quantity unknown. The journey has just begun, and we are all buckled in for the ride.