Search
Choose a style
Dark
Light
Time to read: 6 min

Axom Gaming: where do the biggest AI opportunities lie?

Axom Gaming details how the indiustry can maximise AI usage.
Credit: krungchingpixs / Shutterstock.com

Artificial intelligence continues to cause a significant buzz, and a certain degree of worry in some quarters, but how can you correctly harness this to maximise your ROI?

As technology and player behaviour shift at an ever increasing rate, it’s imperative that businesses exit the experimentation phase with clearly defined strategies.

This is where Axom Gaming comes in. The AI-powered player intelligence platform strives to unlock additional value for operators, while circumventing needless decisions.

In conversation with SBC News, CEO Danny Rippon examines the current state of affairs, discusses where the major problems currently are, details how to avoid giving away more than you need to and much more.

To begin, what would your assessment be on the industry’s current use of artificial intelligence?

I think the industry has made good progress with AI, but there is still a significant gap between what the technology can do and how it is being used today.

Machine learning has been used in gaming for many years to predict player behaviour, identify risk and support commercial decisions. More recently, large language models have broadened the conversation, with operators exploring how AI can improve internal efficiency and reduce manual workloads.

Both have value, but I think the biggest opportunity is in connecting intelligence to action. 

Operators have access to enormous amounts of player data and increasingly sophisticated predictive models, but knowing what a player is likely to do is only useful if that insight leads to a better decision.

That’s where I believe the industry needs to focus. Rather than simply using AI to analyse behaviour or automate existing processes, operators should be looking at how it can improve individual player decisions and deliver measurable commercial results.

At Axom Gaming, that’s the approach we’ve taken. Our machine learning evaluates available actions, expected responses, and incentive costs to select the action and timing most likely to deliver against the operator’s strategy, including taking no action. Those decisions are delivered through the operator’s existing CRM.

We’re also building on this with conversational and agentic capabilities that make it easier for teams to turn their commercial objectives into practical actions, while retaining control over the strategy and boundaries within which AI operates.

Where do you believe, or have witnessed, companies getting their strategies wrong, and how can this be rectified?

One mistake is starting with the technology before defining the problem. ‘We need an AI strategy’ can quickly become a collection of projects without a clear commercial objective.

Another is assuming a better prediction automatically produces a better result. Knowing who is likely to deposit is useful, but offering those players a bonus could simply subsidise something they were going to do anyway.

The starting point should be a specific decision the business wants to improve, such as how to support early engagement or which retention action to take. 

Define success, establish how it will be measured, and give the technology a practical route into the existing workflow. That makes it much easier to distinguish useful progress from an impressive demonstration.

When it comes to maximising ROI from the use of AI, what do you believe are the best strategies to be deployed?

Start with decisions that happen frequently, have a measurable financial impact and can be improved without a major operational overhaul. Small improvements across a large player base can add up to substantial value.

Then measure incremental performance against a comparable control group receiving the operator’s existing approach. Campaign revenue alone does not tell you what AI contributed. 

You need to understand what would have happened otherwise, accounting for incentive costs, technology fees and the effort required to operate it.

Integration matters too. Keeping the existing CRM and adding intelligence to its decisions can shorten the route to value.

Finally, keep learning. Player behaviour changes, and the system should use outcomes to improve future decisions. ROI should be something you continuously demonstrate.

Is it fair to suggest that often less is more can be the best approach when it comes to this technology, and why?

Yes. If a player is going to return and deposit anyway, sending them a bonus can reduce your margin without changing their behaviour. You might record a successful campaign because they redeemed the offer, but you have paid for activity you would have received anyway.

AI should help you decide when spending money or contacting someone will improve the outcome, including recognising when to do nothing. You need a control group to test that; redemptions alone cannot tell you whether the intervention was worthwhile.

‘Less is more’ should be a finding, though, rather than a blanket strategy. For some players, an incentive will generate enough additional value to justify its cost. The job is to distinguish between the two.

Another key talking point often surrounds the hidden cost of bonusing, and the possibility that you are giving away more than you need to. What protocols does Axom have in place to prevent such occurrences?

A bonus being redeemed does not, by itself, demonstrate that it created value. The player may have returned anyway, or the additional activity may not justify the incentive cost.

Axom Gaming evaluates the available actions at an individual player level, considering predicted behaviour, expected response and incentive cost against the operator’s commercial strategy. Those choices can include different incentives, an action without a bonus, or no intervention.

The operator defines the available actions and the boundaries within which Axom makes its decisions. Outcomes then feed back into future decisioning, while comparison against a control group helps establish whether the overall approach is generating additional NGR after bonus costs.

The objective is to direct bonus spend where it creates additional value, while avoiding unnecessary incentives.

As we are quickly hurtling towards the end of the year, what can the industry expect to see from yourselves through 2027?

Our focus through 2027 is to make individual decision-making more autonomous, explainable and easier for operators to use, while adding ‘what if’ scenario planning capabilities. 

We are building on our co-pilot capabilities with a conversational interface and agentic tools that help teams move from a commercial objective to practical action, with less manual configuration. We are also developing the ability for operators to work with Axom through their preferred AI assistant.

For CRM teams, that means more time developing compelling player experiences and setting strategy, with Axom handling more of the individual decisions about which eligible player should receive which action and when.

Alongside that, we want to expand our CRM and platform partnerships, making Axom available within the systems operators already use. The measure of success will remain commercial: better decisions, more efficient incentive spend and demonstrable additional NGR, with operators retaining control of objectives and boundaries.

Subscribe to our newsletter