For years, grocery tech was built for the predictable aisles of center store and retrofitted for produce — usually with messy results. Now, Afresh is flipping the script. After mastering the chaos of the fresh perimeter, CEO Matt Schwartz is expanding the company's AI platform to every department in the enterprise, proving if you can solve for delicate raspberries, you can solve for anything.
With Afresh's expanded platform, grocers can now manage replenishment, demand forecasting, inventory management and distribution center buying across every department on a single grocery-native AI platform.
What does this mean for the retailer's bottom line? To learn more, The Packer recently connected with Schwartz.
The following has been edited for length and clarity.
Traditionally, grocery tech has started in center store and then adapted to fresh. You did the opposite. Now that you're moving into the packaged goods space, what is a lesson learned from produce that is making your center store AI better than traditional systems?
Schwartz: Produce is one of the hardest environments in the store, and it forced us to build a system that can make accurate decisions even when the input data is unreliable.
That's the first lesson: grocery data is messy. If you take it at face value, you'll make bad decisions.
In fresh, that's obvious. There are no barcodes, variable pack sizes and demand that shifts with weather, quality and seasonality. In center store, the data looks cleaner on paper, but the same issues still show up: inventory drift from shrink, mispicks and execution errors at the shelf.
The second lesson is that static rules don't work. A single forecast and a single inventory number aren't enough when both are often wrong. You need a system that accounts for uncertainty. What if demand is higher? What if inventory is lower? And [the system] still makes the right decision.
Most systems were built the opposite way, [with] clean data assumptions, rigid logic [and] one answer. In reality, inventory is often only 50% to 60% accurate, and those systems break as soon as that assumption is off.
We built for that reality in fresh — making decisions that hold up even when the inputs are wrong.
That's why it translates so well to center store. The variability is lower, but the data is still imperfect and the same approach produces better decisions.
Does managing replenishment, demand forecasting, inventory management and distribution center buying across every department on a single grocery-native AI platform improve a retailer's profitability? If so, how?
It improves profitability by reducing coordination gaps and eliminating manual overhead.
Today, fresh and center store run on separate systems, with different data, different forecasts, different order recommendations. No one really has a clear view of how those decisions interact, especially when it comes to cross-merchandising. When everything runs on one system, that changes.
A simple example is promotions. A retailer runs a “family dinner for $15” meal deal that includes pasta, sauce and a baguette. There's no
produce in the bundle, but customers add salad, tomatoes or a cucumber.
Without visibility, produce reacts after the shelf is already light. With it, the system adjusts ahead of time, and just as importantly, brings orders back down when the promotion ends, without someone having to go in and fix it.
The other piece is overhead. Legacy systems require teams of analysts to constantly tune forecasts for seasonality, promotions and holidays. That's ongoing maintenance. With our system, that work largely goes away. The models update as conditions change, so corporate teams can spend more time on strategy and less time tuning forecasts.
As Afresh expands beyond fresh to include everything from shampoo to frozen pizza, how will you ensure continued innovation focused on fresh produce and other fresh items?
Afresh is a fresh-first AI company, and that doesn't change. We started there because fresh is the most complex part of the store — and it's still where we push the hardest. Expanding to center store isn't a shift away from that. It's the same system applied more broadly.
The core problems we solve — bad inventory, demand variability, execution gaps — exist in every department. What changes is the context.
When more of the store runs on the platform, fresh decisions actually get better. The system can see promotions, traffic patterns and how customers shop across the full basket.
This isn't about becoming a general retail system. It's about extending a fresh-first architecture across the store in a way that continues to improve fresh.
You mentioned a 95% or more adherence rate. How will produce managers benefit from using a unified platform with other department managers? How does this improve or change the daily workflow for a produce clerk?
Our 95%+ adherence rate means store teams trust Afresh across all departments.
For produce managers, the workflow doesn't fundamentally change, but the quality of the recommendation does. Now those recommendations reflect what's happening across the entire store, not just produce. That leads to fewer misses with fewer out-of-stocks and less shrink.
There's also a labor benefit. When all departments run on the same platform, teams can flex more easily. If a produce manager is out, someone without years of department-specific experience can still place a good order because the system carries that historical context.
And at the corporate level, produce performance is no longer isolated. It's visible alongside every other department, which drives more consistent focus and investment.
How does having every item in the store on one platform specifically help the produce department? Are there additional benefits to cross-merchandising strategies throughout the store here as well?
Today, produce teams see an item sell faster than expected and have to figure out why. That delay is what leads to missed orders and empty shelves. When every item in the store runs on the same system, you can see what's driving that demand.
For example, if a retailer features ranch dressing on an endcap, more customers buy carrots and celery to go with it. Without that visibility, the system just sees carrots selling faster than expected, and stores react after they've already sold through. With it, the system knows the promotion is happening and increases orders ahead of time.
The same thing happens with substitutions. If a packaged item goes out of stock, customers shift into fresh alternatives. Without that signal, produce teams chase the demand late. With it, they adjust early and, just as importantly, reduce orders again once the packaged item is back in stock.
Can you share an example of how a produce buyer at the DC makes a better decision because they now have visibility into the inventory levels of the non-perishable aisles?
A simple example is a short-term supply issue in center store. Let's say a top-selling jarred salsa is out of stock in a region for 10 days.
That doesn't always change behavior, as customers may switch to another jarred option. But in some cases, you do see a small shift into in-house made pico or fresh alternatives.
The challenge for the produce buyer is knowing whether that change is real and how long it's going to last. Without that context, they might see a slight lift and either ignore it or overreact. With visibility, they understand the cause. That allows them to make a small, measured adjustment and, more importantly, to unwind it at the right time.
For fresh items with a 3- to 5-day shelf life, timing matters more than the initial increase. The risk isn't missing a bit of upside; it's being left with excess inventory after the demand disappears.
Afresh is known for reducing food waste. How will expanding to center store and general merchandise help retailers to reduce waste throughout the store?
Afresh has prevented more than 200 million lb. of food waste since founding, and most of that has been in fresh, where waste rates are highest. But the underlying problem is the same across the store.
In fresh, waste shows up quickly as spoilage. In center store, it shows up more slowly through excess inventory, markdowns or discontinued products that don't move. In both cases, it comes from ordering too much or ordering at the wrong time.
When forecasting improves and inventory is more accurate, products move through the system faster. That means less spoilage in fresh, and less excess and fewer markdowns in center store.
Do you have a retailer success story you can share from the produce perspective? How do you see this success evolve as Afresh tech expands to the whole store?
Across our retail partners, we consistently see the same pattern in produce: shrink comes down, sales go up, inventory turns improve and store teams adopt the system quickly because it earns their trust. In many cases, that means double-digit shrink reduction, sales lift and meaningfully faster turns within the first few months.
As the platform expands across the store, that impact compounds. Retailers don't have to rip out their existing systems or go through a new, heavy implementation to get there. They can expand incrementally, adding new departments onto the same platform or integrating Afresh as the decision engine behind their existing tools.
In practice, that means the store teams don't have to learn a completely new system, and corporate teams don't have to rebuild their tech stack.
