Xtremepush: heightening quality and improving efficiency

Artificial intelligence has risen in prominence to become a mainstay in many iGaming conversations. In recent times, it could be argued that this is only rivalled by Brazil as being the most popular buzzword of the moment.

In the concluding part of an Xtremepush special, the customer engagement platform looks beneath the surface of a technology that is widely lauded as boasting the potential of transforming the entire iGaming landscape.

Despite promising significant possibilities across a range of segments, it is around critical topics such as personalised experiences and player engagement where Xtremepush is primarily interested.

However, before diving into the nitty gritty, Tommy Kearns, CEO and Co-Founder, offers SBC News a little perspective on the group’s current usage.

“At the most basic level, machine learning has been around for 30 years or more, but we’ve been using this for the last three or four years,” he says. “What we’ve been using those algorithms for to predict players behaviours such as churn, becoming a VIP etc.

“We are building AI into our support system more and more”

“Looking at VIPs, better experiences, cross sell and segments to predict the best next action. That’s how it helps in a general sense. 

“There are plenty of use cases. The really important ones centre around churn prediction, VIP, prediction experience or cross sell prediction, cohort analysis, recency and frequency monitoring, and large language models.”

Continuing the conversation, Morten Pedersen Tonnesen, Chief Growth Officer, first touches on the “very basic internal” usage, which centres around utilising generative agents (chatbots) for content creation. 

However, an internal uptake is expected to be ramped up moving forward when it comes to assisting in building individual relevant experiences for players.

“We are building AI into our support system more and more,” Tonnesen comments. “This includes building into our generation of content, enablement and everything else. 

“From an internal standpoint, we are doing a lot in terms of heightening quality as well as improving efficiency, which basically means getting more of a higher quality out with less people.

“Of course this is a key parameter in terms of staying competitive and keeping the cost base as manageable as possible. 

“When it comes to our product itself, what we are doing there is that, the core of any CRM platform has been capturing all the data and then utilising the journey builder to build different treatment strategies for different cohorts of players and individual players.”

“We’re not replacing them. That’s not going to happen. That’s been a faux path”

Regarding the path that lies ahead, reference is swiftly made to that of another popular industry buzzword – innovation. 

This regards work undertaken during the past six to 12 months on ‘the agentic layer’, which essentially concerns augmenting an agent with your CRM professional. 

“We’re not replacing them,” Kearns reassures. “That’s not going to happen. That’s been a faux path for a long, long, long time, but what we are seeing is CRM execs using large action models, or the agent layer, to help them in their business on a day to day basis. 

“Automatically analysing cohort analysis and then bubbling up the next best action and then literally doing it for them.

“All that the CRM executive has to do is say, yes, I do like that cohort now. Yes, that makes sense, and push to approve it. It’s massively enabling, much more efficient for CRM execs and teams, and also allows more output at scale.”

However, Kearns warns that this can only be effective when built upon solid foundations. Xtremepush’s foundational layer must be plugged into operators “before they even start to talk about the agentic layer”.

In simple terms, this agentic layer can only work on top of its foundational counterpart as that is how it gains the data, and subsequently an understanding of what they should be doing.

“When you marry that real time foundational data, decisioning, consent regulatory requirement layer to the agentic layer, that’s where the magic’s about to happen,” he concludes.

“The vision is over the next couple of years that the journey builder becomes more and more automated”

“You’re meeting the players where they are, you are actually automating it at scale, being exceptionally helpful and creating experiences that create loyal players which drives life time value.

“It’s very, very early for that in the industry. We have three or four agents that we’re currently bringing to life, like CRM executive augmentation, an Agent Analyst to help understand and drive the next best action, all of those kinds of things. That’s going to be phenomenally exciting within this industry over the coming next 12 to 24 months.”

Concluding his own participation, Tonnesen promises that the coming months will witness the launch of a range of agents that can handle different parts of the customer journey. 

“The vision is over the next couple of years that the journey builder becomes more and more automated and you will need less people to sit and drag and drop and build journeys for players,” he sums up.

“It will become the AI that figures it out. One customer has certain specific needs compared to customer two and so on, and we’ll continuously optimise this via pulling dynamic content from a CMS, pulling real-time one-to-one offers and communicating, changing the product and everything within the boundaries of what’s set by the CRM team.”