Groceryshop 2026 recap: What grocers told us about AI, data, and fresh

My favorite part of any trade show is sorting signal from noise. At Groceryshop this year the signal was unusually loud. In meeting after meeting with grocers, wholesalers, distributors, and technology partners, people in completely different roles described the same problems, and almost every conversation started with AI. Here are the themes we heard most often, and what I think they mean for grocery.

Building agents is the easy part of AI. Having good data underneath it is the challenge.

Nearly every conversation about AI turned into a conversation about data and trust. One grocer told us about an old cauliflower UPC that got reused for a floral item. Another inherited a newly acquired banner whose data was too messy to run on the parent company's systems.

The comment that stuck with me most came from a digital leader. Even with good governance, he said, fresh data goes stale fast. Pack sizes change, local vendors come and go, and codes get reused every week. Talk about data attribution nightmares.

That's why data readiness in fresh has to be continuous. The cleanup has to run inside the intelligence layer, all the time, or the agents will quickly get off track. We built our solutions to do that harmonization as part of the daily work, so grocers can start now and let data quality improve as an output of running our platform.

The best solution is the one people actually use

Once the conversation moved past data, it landed on people. Grocers cared much less about how sophisticated an AI model is than about whether their teams, often tens of thousands of associates, will trust it and use it.

That trust gets won or lost on the floor. A department manager remembers one bad order for months. A manager who gets a good order every day stops second-guessing the system and spends that time on the display, the customers, and the team.

Keeping people in control is part of earning that trust, but control can't mean reviewing everything. A produce manager doesn't have time to check every line on every order, and they shouldn't have to. The model that works is exception-based. The system handles routine items automatically and uses AI to flags the few that need a human's judgment, like a new vendor item, or an inventory count that doesn't match recent sales.

Over time, the team spends less time checking the system's work and more time on the work only people can do: talking to customers.

The opportunity for solutions that solve real-time inventory intelligence is growing

One of the biggest opportunities we saw was the race to understand what is actually on the shelf at any given moment. Technology providers are coming at the problem from every angle: advanced cameras, shelf-scanning robots, optical intelligence, RFID, and inference models that combine multiple signals into a real-time picture of inventory. Until recently, most of that work focused on center store, where products are packaged, labeled, and relatively easy to identify.

Now that intelligence is moving into fresh. The problem is harder, but the potential value is greater. New solutions are starting to estimate how full a produce display is and to pick up signals about quality and freshness, going beyond whether an item is simply present. That richer view of the shelf could help grocers replenish sooner, protect the customer experience, and give forecasting and ordering systems a much better picture of what is available to sell.

It also brings us right back to data. Every camera, sensor, and scan is one more signal that has to be cleaned, reconciled with everything else the grocer knows, and turned into a decision a department manager trusts. Shelf intelligence pays off when it feeds an intelligence layer that can act on it.

What this adds up to

Across these conversations, grocers are thinking about AI in exactly the same way: how do I get the foundations and pipes set up for these tools to have real value, and be trusted by my team. That’s the real blocker in AI adoption, and only can be solved with partners who deeply understand the space and are willing to do the work to help you get there — without it taking an entire year.

As for Groceryshop itself, our chickens made the loudest noise on the show floor. That seems fitting, since the rotisserie chicken may be the best example of everything above. It's one of the most loved items in the store and one of the hardest to have ready at exactly the right moment!

Thanks to everyone who stopped by, and we'll see you next year.