At Raving NEXT: Casino Strategy & Operations Summit, one session didn’t just land—it lingered.
AI That Pays: Top Five Integrations Changing Casino Profitability Right Now, led by Andrew Cardno, Chief Technology Officer of Quick Custom Intelligence, sparked a wave of follow-up questions from operators looking to move beyond AI hype and into real-world application.
What came next were sharper, more practical questions—the kind that don’t always fit neatly into a 60-minute session.
So we took those questions back to Andrew and got his take.
The Questions Operators Are Asking About AI Right Now
1. How can AI assist marketing with pre- and post-forma creation and analysis to optimize this important reporting?
For a long time, pre-forma and post-forma analysis followed a relatively linear process. Assumptions were built from historical averages. Campaigns were executed. Results were reviewed. Variance was explained, often after the fact.
The structure was sound, but the feedback loop was slow.
AI compresses that loop.
On the pre-forma side, it allows assumptions to become adaptive rather than fixed. Instead of relying solely on historical averages, models begin to reflect behavioral patterns—how different players respond to timing, reinvestment levels, and offer structures.
On the post-forma side, the shift is from reporting to diagnosis. Not just what happened, but why it happened, at a level of detail that would be difficult to achieve manually.
The real opportunity is when those two processes are no longer separate. When post-forma insights feed directly into the next pre-forma automatically, the reporting function begins to act as an optimization system.
That is where the value compounds.
2. Which specific AI integration has delivered the biggest measurable impact on casino profitability so far, and what made it successful?
The most consistent impact has come from individual-level reinvestment optimization.
Historically, reinvestment decisions have been made at the segment level. That approach is practical, but it introduces inefficiency. Some players are over-incentivized, others are under-engaged.
AI allows that decisioning to move to the individual level.
Reinvestment can be aligned more closely with predicted behavior, expected value, and responsiveness. The result is typically improved margins alongside stronger engagement.
What makes these implementations successful is not just the model.
It is integration.
The optimization must be embedded directly into campaign workflows. It must be trusted by marketing teams. And it must continuously learn from outcomes.
Without that, it remains an analytical exercise.
With it, it becomes operational—and the financial impact becomes visible.
3. How are companies that use AI avoiding copy infringement and plagiarism?
Most organizations that are approaching this seriously are not treating AI output as final content.
They are treating it as a draft layer.
AI is used to structure, accelerate, and refine thinking. But outputs are reviewed, edited, and aligned to brand standards before they are used externally.
In more controlled environments, outputs are also grounded in internal or approved content sources, which reduces the likelihood of unintended overlap.
The underlying principle is straightforward: the responsibility for originality still sits with the organization, not the model.
The companies that run into issues are typically the ones that collapse that step.
4. What are your thoughts on the projected crash of OpenAI in 2027?
Predictions of collapse tend to appear alongside every major technology transition.
They are usually less about the company itself and more about uncertainty around the pace of change.
What is more relevant is that the underlying capability—generative reasoning—has already been adopted at scale. Individuals and organizations are integrating it into daily workflows because it is useful.
That level of adoption tends to be durable.
Individual vendors will compete. Some will perform better than others. But the category is expanding, not contracting.
From a strategic standpoint, it is less important to predict which company wins, and more important to ensure that your organization is not dependent on a single provider.
Flexibility is the safer position.
5. What are some solutions or thoughts on the use of clean water that data centers require to operate?
For many years, compute was treated as an abstract resource—measured in performance and cost, but not always in physical inputs.
That is changing.
Data centers require energy, land, and in many cases, water for cooling. As demand increases, those inputs become constraints.
The response is already evolving.
We are seeing movement toward more efficient cooling systems, including closed-loop and air-cooled designs. We are seeing geographic shifts toward regions where resources are more abundant. And we are seeing increasing transparency requirements around environmental impact.
Over time, these factors will influence vendor selection more directly.
What was once an operational detail is becoming part of the strategic evaluation.
6. Can a company like Raving assign an AI expert to help us write casino marketing-centric prompts?
They likely can, but the more important question is what problem is actually being solved.
Prompt writing is useful, particularly in the early stages of adoption. But it is not, by itself, a long-term capability.
The organizations that see sustained value tend to move beyond individual prompts and toward structured use cases. Campaign design, segmentation logic, host communication, reinvestment strategy—these become repeatable workflows supported by AI.
At that point, the prompt is simply an interface.
If the support being offered helps translate real marketing workflows into consistent, repeatable AI-assisted processes, it will have value. If it remains focused on how to phrase individual prompts, the impact will be more limited.
If the original session focused on what’s working today, these follow-up questions point to what operators are trying to solve next.
This is where AI shifts from interesting… to operational.
Read the full session recap: AI That Pays: The Moment the Curve Collapsed →