Dylan Winik, CEO of Oceania at Nayax, discusses where the greatest value of AI lies in terms of how much distance it closes between analysing data in its simplest form and making a confident decision. He argues CIOs should stop chasing AI adoption as a goal in itself and instead recognise that while AI can provide the signal, humans provide the context and make the decision.
Over the past decade, many organisations have become increasingly good at collecting data. Transactions, inventory, customer behaviour, equipment performance and operational activity can all be measured in extraordinary detail. But in practice, greater data visibility does not automatically make it easier to decide what to do next.
For many businesses, answering a seemingly simple question can still mean moving between dashboards, comparing reports and manually piecing together information before anyone can understand what has changed and why.
That’s where some of the most practical applications of AI are beginning to emerge.
Rather than asking how much AI an organisation can deploy, technology leaders should start by asking where people are currently spending time finding, comparing and interpreting information before they can make a decision.
In many cases, this is the greatest value of AI. It can reduce the workload required to make an informed, data-driven decision and a human judgement call.
Reduce the distance between signal and decision
Many organisations are sitting on valuable operational information but lack the time or resources to analyse it continuously.
An operator might want to understand why sales dropped at a particular location, which products are consistently underperforming, whether stock levels are becoming a problem, or whether an unusual change in performance needs investigation.
Traditionally, answering those questions requires somebody to write reports, compare different periods and work through individual data points before identifying the issue.
Today, AI-enabled tools can increasingly do some of that groundwork. They can summarise information, identify unusual changes and make it easier for teams to understand which areas deserve their attention.
We are seeing that evolution within Nayax, having recently introduced an AI assistant that allows operators to ask questions about their business data in plain language rather than manually constructing reports. Additional AI capabilities analyse historical sales information to recommend changes to product mix.
The outcome that changes here is the shift between having information and knowing what to investigate or do next.
Make recurring operational decisions easier
Some of the strongest use cases for AI are likely to be the recurring decisions organisations already make every day.
Inventory is one example. Businesses need to understand what is selling quickly, what requires replenishment and what may be sitting idle for too long. If those patterns can be surfaced automatically, employees can spend less time reviewing every metric and more time responding to the changes that matter.
The same principle applies across broader operational performance. Instead of expecting people to monitor every data point equally, technology can help highlight exceptions, anomalies or emerging trends that warrant closer attention.
That becomes particularly valuable in distributed operating environments, where businesses may be managing large numbers of devices, locations, payment points or other customer touchpoints.
The challenge for CIOs is therefore not simply collecting more data; it’s designing systems that make existing data more useful to the people responsible for acting on it.
Automate the analysis, not the accountability
There is an important distinction between identifying a pattern and understanding it.
An AI system may flag that transaction volumes at a particular location have fallen sharply. It cannot necessarily know all the variables: whether local roadworks have reduced foot traffic, whether a machine has been moved, whether customer behaviour has changed or whether another business has opened nearby.
Similarly, a system might identify increased demand for a particular product. Someone working closer to the business may know that it is connected to a promotion, seasonal change or a shift in customer preferences.
AI can provide the signal, but people provide the context and make the decision.
That distinction matters as organisations look for opportunities to automate more business processes. The objective should not be to remove human judgement wherever possible; the goal is to remove unnecessary work around that judgement.
If technology can reduce repetitive reporting, information gathering and manual analysis, employees can spend more time on decisions that require commercial knowledge, customer understanding and experience.
Measure outcomes, not AI adoption
One of the risks of the current AI cycle is that organisations begin measuring progress by how many tools they have introduced.
‘Using AI’ shouldn’t be a business objective in itself. A more useful test is whether the technology makes a particular process meaningfully better.
Before adopting an AI capability, leaders should ask:
- Which specific problem will it solve?
- Will it remove work from an existing process, or simply add another platform to manage?
- Can employees actually act on the information it produces?
- Is the underlying information reliable?
- Where should human judgement and accountability remain?
- How will we know whether technology has improved an outcome?
The quality of the underlying information is particularly important. AI is only as useful as the information it draws from. A sophisticated interface cannot compensate for poor, fragmented or unreliable data.
The same applies to usability. Generating more insights is not inherently valuable if they create another queue of information for employees to work through. The goal should be to make decisions simpler, faster or better informed.
Start with the friction you already understand
Organisations should plan carefully when introducing AI across the business.
The most practical starting point may be a recurring frustration that employees already understand well. Perhaps a report takes too long to prepare, teams spend hours comparing information across systems or operational problems are noticed only after they have already affected sales or customers.
Those are concrete problems against which AI can be tested.
Ultimately, the best measure of AI is not how sophisticated the technology sounds, but whether it reduces the time required to answer an important question, helps teams identify a problem earlier or gives people more capacity to focus on higher-value work.
For CIOs, that may be the most useful way to separate genuine operational value from AI for AI’s sake.
The organisations that get the most from the technology will not necessarily be those deploying it as much as possible, they’ll be the ones using it to shorten the journey from data to insight to a better decision.
