Personalisation is hardly a new ambition for the gambling industry. For the best part of two decades, CRM teams have been slicing databases into progressively smaller groups, adding more data points and building increasingly sophisticated models of customer behaviour. For Anthony Smith, the important question is no longer how precisely an operator can segment its players. It is whether it still needs to segment them at all.

Smith joined EveryMatrix after spending much of his career on the operator side, where he had experienced first-hand the limitations of the technology available to marketing teams. Big data gave those teams far greater visibility of their customers, but it did not necessarily solve the problem of scale.
“You could create an infinite amount of segments, but you then had to bring them back down to a manageable level,” he explains. “You still weren’t achieving one-to-one.” AI, he believes, changes that equation. “The thing I get very passionate and excited about is that suddenly we are able to do true one-to-one personalisation, because essentially every creative touchpoint can be designed and manufactured in real time.”

Touchpoints could mean an email, SMS or push notification, but Smith is interested in what happens while the player is using the product. Rather than changing a name or offer within a pre-built template, the interaction can potentially be created around the individual player and what they are doing at that moment. It’s a significant change from the conventional CRM model, but Smith stops short of calling it a revolution. “I think it’s still incremental, but I do think it’s unlocking opportunity,” he says. “And it’s really happening.”

When the moment has already gone

Smith describes the traditional operator experience of sitting in a Tuesday trading meeting, identifying something unusual in the weekend’s performance and sending analysts away to investigate. Three or four days later, the answer comes back. “The problem with both the consumer experience and the product is the moment’s gone,” he explains.

For larger operators, the difficulty can be compounded by data arriving from numerous suppliers and different systems before it can be reconciled into a single customer view. EveryMatrix’s argument is that its native architecture removes much of that process because the underlying sportsbook, casino and customer data already sits within the same environment.

For Smith, however, speed alone is not the breakthrough. “You can be instant but still really, really irrelevant, or you can use AI and big data to be super relevant but be too late for it to matter. It’s not one or the other. It’s both.” The difference becomes apparent during live betting. Previous systems Smith worked with could potentially respond game by game or session by session. EveryMatrix is working towards responding to individual events within the game.

A goal, card or other event could, therefore, become another contextual signal from which the system determines the appropriate next action, which changes CRM from something that predominantly analyses what happened yesterday into something capable of becoming part of the experience taking place now. Smith sees an opportunity to make betting more engaging rather than simply extracting greater spend.

“If we’re not always trying to stretch spend per head, we give you more entertainment for that money.” The upshot might mean another relevant market, a real-time bonus or simply a different experience based on what has happened during the event. The objective is not necessarily to persuade a customer to spend £30 today if doing so damages a relationship that could otherwise generate £10 a week over a much longer period.

Real-time intelligence still needs a longer-term view of the player. “That tension between real time, making it relevant and enriching the betting experience, as well as that long-term lens – they complement each other” underlines Smith.

Just because you can…

If technology makes it possible to contact customers more frequently and with ever more sophisticated offers, personalisation could simply become a more efficient form of spam. To counter this, Smith has a simple rule. “Just because you can doesn’t mean you should.”

Relevance, he argues, has to extend beyond understanding what somebody is betting on. It also needs to understand the broader relationship with that player, including their activity, previous behaviour and potentially signals suggesting that the correct action is not another promotional message at all. This becomes particularly important as traditional marketing channels lose effectiveness.

Smith says email opt-in rates that might once have reached 60 or 70 per cent can now sit at 20 or 30 per cent in markets he’s worked in. In which case it’s the platform itself that becomes increasingly important.
“When you’re on platform, that’s the time to make a difference,” claims Smith. Rather than trying to pull a player back through another email, CRM becomes embedded within the product. The sportsbook or casino interface can respond to what the customer is doing. It also explains why Smith repeatedly returns to one word: context. Adding a large language model to a CRM platform is relatively straightforward. Making its output commercially useful, consistent and compliant is considerably harder.

EveryMatrix is developing what Smith describes as “context libraries,” establishing what the AI should understand about the customer, the operator’s objectives and the regulatory environment before determining an action. Different markets impose different restrictions. Customers have different permissions and risk profiles. Operators have different commercial priorities. An AI system capable of generating almost unlimited variations needs to understand those boundaries before its speed becomes useful.

“Marketing has always been the search for context,” Smith says. “These tools don’t change that. If anything, they bring it forward because you are now able to do so much more.” EveryMatrix is consequently resisting the temptation to race AI functionality into production simply because it can. “We’re not rushing to be first mover in this space. Particularly in this area, it’s about getting it right and making sure that it’s adding value.”

Humans remain in the loop

Smith does not envisage CRM becoming an entirely automated loop in which the machine identifies the customer, decides the objective, creates the interaction and executes it without human involvement.
The location of that human involvement may change considerably, however. Instead of manually producing individual campaigns, teams can establish business objectives, regulatory requirements and compliance rules, allowing technology to perform much of the repetitive work underneath them.

“I don’t ever foresee a day where you haven’t got human oversight,” Smith says. And there’s another reason for keeping people involved. While AI can analyse, automate and generate at extraordinary speed, Smith questions whether access to the technology alone will provide lasting competitive differentiation as increasingly capable models become widely available.

“The technology is incredibly powerful, but ultimately operators still need to decide what they want to do with it,” he says. “AI can analyse data, identify opportunities and help execute ideas at a scale that simply wasn’t possible before. But the commercial thinking – understanding your customers, deciding where you want to differentiate and coming up with the ideas that set you apart – still needs a strong human element.”

Smith describes his experience of AI at EveryMatrix as “an assist, not a replace.” It may reduce the number of people required to perform certain tasks and alter the skills demanded of CRM teams, but that is different from eliminating them. The opportunity is to remove the layers of manual execution between an idea and the customer.

Democratising CRM

Sophisticated CRM has traditionally required sizeable teams of analysts, marketers, creatives and specialists. Smith recalls working with teams of 30, 40 or even 50 people across major operators and multiple markets. Scaling meant adding people – while AI potentially reverses that relationship.

“I want to democratise the product and put it in a commercial team’s hands,” he says. Traditionally, a sportsbook or casino team might establish its commercial objective before handing it to CRM specialists, who translated that objective into segmentation, campaigns and execution. Results would then make their way back through the organisation before the next round of optimisation. Smith wants technology to remove much of that “middle jungle.”

The aim is not to remove expertise from the system but to build it into the system, putting sophisticated capabilities directly alongside the people making commercial decisions.

For smaller operators, the implications could be substantial. Capabilities that previously depended on the resources of a tier-one operator could become available without replicating its headcount.

EveryMatrix is already testing that proposition in the market. Its native CRM is live with a UK-regulated customer, while its broader Growth Services proposition is being used to identify opportunities within operator data. Smith said analysis conducted with three partners had identified potential revenue opportunities of between 20 and 30 per cent, although the next challenge is converting those insights into action.

For Smith, this is where CRM, managed services and the wider EveryMatrix platform begin to intersect. An operator can identify an opportunity through its data, determine an action and execute it through the same underlying ecosystem, with managed services available where it does not have the internal resources to do so itself.

One player, one journey?

Does all this mean the segment is ultimately heading for extinction? Smith is understandably reluctant to make predictions about technology moving at its current pace, but he can envisage segmentation becoming much less important to the customer experience. “I can foresee a day where true personalisation, the experience, is real one-to-one.”

Big data already made it theoretically possible to create extraordinarily granular customer groups. The limitation was operational. An operator could create a million segments, but it could not realistically design and execute a million different customer journeys. AI begins to remove that constraint. However, that’s not the same as saying segmentation will disappear entirely. Operators still need to group customers to understand trends and report on performance. Nobody wants to examine a dashboard containing a million individual lines.

What’s more likely a scenario is that the segment may cease to dictate what the player actually sees. Instead, every player could effectively occupy a segment of one, with their experience shaped continuously by their behaviour, preferences, circumstances and the context of that particular moment. Whether consumers want that degree of personalisation will determine how far the technology goes.
For Smith, that’s the test that matters. “If it doesn’t add value to the betting experience,” he says, “why do it?”

After years in which CRM has promised increasingly sophisticated personalisation while remaining constrained by the practicalities of segmentation, that question may prove more important than the technology itself. EveryMatrix now has the data, AI and infrastructure to get much closer to the individual. The challenge is knowing when it should.