RetailNext CRO Perspective: The Next Two Years Of Retail Analytics

Earlier this year, we hosted our annual Executive Forum. Close to 100 retail leaders spent three days with us, spanning breakouts, dinners, and unscheduled hallway conversations. One conversation came up in every one of those settings. It started with AI. It pivoted, almost always, to a harder question: what does any of this mean for the analytics infrastructure we already have?

Watch The Session 👉 How Foot Locker Is Using Store Data To Drive Conversion

The Conversation Retail Leaders Are Actually Having

The board-level AI conversation in retail has matured fast. Two years ago it was, "should we have an AI strategy?" One year ago it was, "where are we piloting?" Today it is, "which parts of our operating stack hold up under what we are trying to do, and which parts do not?"

That last conversation is harder than the first two. It forces leaders to look at infrastructure they bought years ago, often through procurement, and ask whether it is still fit for purpose. For most physical retailers, traffic and store analytics are exactly the kind of infrastructure that sits in that category. Bought once. Largely working. Quietly limiting.

The money tells you how serious this has gotten. Gartner forecasts that global AI spending will hit $2.52 trillion in 2026, a 44% year-over-year increase. IBM surveyed 1,500 global retail and consumer products executives and found that 81% are using AI to a moderate or significant extent, with their internal teams running even further ahead. These are not pilots anymore. These are operating budgets.

Why Legacy Counting Infrastructure Cannot Carry It

The legacy traffic counters that have anchored this category for more than three decades were built to answer one question: how many people walked in the door? At the time, that was useful. Conversion rates, staffing ratios, and hourly trends gave operators a real edge.

But the ceiling of that architecture is hard. Hardware that reports a number cannot become software that learns from patterns. AI capabilities bolted onto a counting platform are a structural mismatch, not a product update. The data foundation is not there.

The retailers who recognize this early will modernize their analytics infrastructure as part of their AI strategy, not after it. The ones who do not will spend the next two years pouring AI investment into a foundation that cannot support the weight.

I have watched this pattern play out in adjacent categories. The companies that get to AI native operations first establish a structural advantage. The companies that retrofit AI onto legacy stacks spend more, move slower, and produce thinner results. The gap is real, and it widens fast.

There is a deeper question behind all of this. The analytics platform a retailer chooses for the next decade needs to support AI use cases that have not been invented yet. The only way to deliver that flexibility is to architect machine learning into the platform from the early years, not bolt it on once the market demands it. Retailers should be asking their providers a different question. Not just what does the platform do today. But rather: How was it built? Legacy traffic counters were not built for this. The platforms that win the next ten years will be.

What AI-Native Operations Actually Look Like

Most of the category is still selling and buying weekly reports. A smaller set of retailers has already moved past that. Here is what that move looks like in practice.

• A US department store has RetailNext device health signals auto-feeding ServiceNow, so incident tickets open and close without human intervention. Operational alerts that used to require someone to notice and route are now handled by the workflow itself.
• A Canadian premium apparel retailer is building its own analytical models on top of RetailNext data and distributing the outputs across digital, marketing, real estate, and operations. The platform is the foundation. The intelligence layer they have built sits above it and crosses every function in the business.
• A US athletic retailer is using AI to generate tasks. When a metric is off in a given store, the system flags it, generates the action for the district manager to take, and tracks whether it gets resolved. The team is treating store performance as a continuous closed loop, not a weekly readout.
• A UK specialty athletic retailer is replacing a quarterly manual labor rota process, run by one person across hundreds of stores, with continuously updated, predicted traffic-driven labor allocation. The planning cadence is moving from once a quarter to once a week.
• A US home furnishings retailer is replacing an inherited people counting system with a live API feed and AI-assisted forecasting, with their internal analysts brought directly onto the platform.

Five different retailers. Five different categories. The pattern is consistent. Continuous instead of weekly. Integrated instead of isolated. Forward-looking instead of backward-looking. The retailers operating this way are not waiting for the category to catch up. They are pulling ahead while their competitors are still in evaluation cycles.

What The Next 24 Months Will Look Like

Three things will be true two years from now that are only partially true today.

1. Legacy traffic counting infrastructure will be visibly obsolete at the enterprise tier. The retailers who have not modernized by then will be operating with categorically less information than the ones who have. Not slightly less. Categorically less.
2. AI-native platforms will have moved from a competitive advantage to a baseline expectation. The retailers who built early will compound the lead. The retailers who waited will spend the next two years catching up to a position the leaders are already moving on from.
3. The cost of getting this wrong will become visible in financial results. Labor productivity, conversion performance, real estate decisions, marketing ROI. The gap between AI-native operators and everyone else will be measurable on the P&L. That is the moment when most boards stop debating the timeline.

The Decision That Cannot Wait

If you are a retail leader reading this, the question to take into your next strategy review is not "are we doing enough on AI?" That question is too abstract to be useful. The question is: does the analytics infrastructure we have today support the operating model we want to be running in two years?

If the answer is yes, you are ahead of most of the category. If the answer is no, the cost of waiting compounds every quarter. The retailers moving fastest right now have already decided.

The category is not standing still. The question is whether your infrastructure is moving with it.

Another Session? 👉 Kendra Scott & Marine Layer Are Turning Stores Into Community Hubs