2025 continues to be a year of ups and downs. Inflation has lessened slightly, and seasonal spending in the UK is expected to increase +4% to £26.9billion this year. However, food prices remain stubbornly high and consumer confidence has dipped. And that means grocery retailers face the unenviable challenge of controlling prices and delivering value to satisfy an increasingly wary consumer, says Kevin Johnson, CEO of Focal Systems.

Festive shoppers expect stocked shelves every year, but the stakes are likely to be even higher for retailers at a time of mixed consumer sentiment.

Stores must meet expectations for shoppers ready to spend, while avoiding giving any customers reason to pull back – particularly those worried about economic conditions. Grocers, in particular, cannot afford to have any gaps on-shelf. During the holidays, consumers are especially unwilling to make substitutions. One out-of-stock item can lead to a completely abandoned basket – particularly when shoppers are wary of overspending.

To avoid frustrated customers and lost sales, retailers must continually monitor product availability and employ a reliable, efficient process to ensure timely and accurate replenishment. Fortunately, the latest advances in AI, such as computer vision shelf monitoring, enable retailers to do what was previously impossible: increase on-shelf product availability while controlling operating costs. “Shelf AI” can help retailers in multiple ways, including increasing inventory accuracy, reducing out-of-stock instances and duration and redeploying labour from low to high value tasks.

Continuous monitoring, greater accuracy

The first step in keeping products stocked and available is knowing what is on-shelf and in inventory within each store. That’s far harder than it seems thanks to supply chain challenges, shrinkage, the cost of labour and even human error.

Until the advent of Shelf AI, the standard process of checking product levels required store workers to walk around the store and manually record what is fully stocked versus low or out, either with pen and paper or with a handheld scanning device.

Such a system is labour-intensive, typically requiring many hours to scan a single store just once per day. Unfortunately, due to human nature, labour shortages and customer interruption, such a system is also prone to error, leading to an inventory accuracy rate of only 65% in a typical grocery store, making timely and accurate replenishment essentially impossible.

When grocers use AI-powered shelf cameras to determine inventory levels, they no longer rely on once-daily aisle walks. Instead, strategically placed cameras can track product availability every hour of the day. Shelf AI can also provide retailers with much more accurate product data – 95% instead of 65%, keeping inventory data aligned to the demands of the season.

Reducing out-of-stocks, increasing availability

With shelf cameras, inventory and availability numbers are not only more accurate, but far more timely.

When humans scan store shelves only once per day, it can take as long as 24hours for Order Management Systems (OMS) and stockers to know that a product is out-of-stock. By the time it is restocked from existing inventory or reordered and then restocked, a product can be unavailable and therefore “off sale” for much longer than 24 hours.

With Shelf AI delivering hourly scans of every product on every shelf, store workers can be alerted of low or out-of-stock situations within 30 minutes, leading to action within the hour.

Timely replenishment resulting from computer vision can keep products from going out-of-stock at all, meaningfully boosting product availability. And if a product does go off-sale, it can be reordered and refilled hours and even days earlier – saving thousands of lost sales and disappointed customers per store.

Helping customers instead of hunting for products

Additionally, by using cameras to scan products instead of humans, retailers are able to shift labour hours from low-value menial tasks to value-adding, high-impact work.

Every hour store associates spend walking the aisles scanning out-of-stocks or hunting for inventory is an hour they cannot spend putting products on shelves or directly assisting customers.

By deploying Shelf AI, the average grocer can shift 300 labour hours per month from walking and scanning to active restocking – further boosting product availability, sales and customer satisfaction.

This year’s Peak Trading season may be more challenging for retailers than most, but stores that have already deployed computer vision to help manage replenishment will be better positioned than stores still operating without the assistance of Shelf AI.