GNC, a health supplements company, has sharply improved inventory accuracy and greatly reduced unshipped orders caused by stockouts after introducing artificial intelligence (AI) drones at its logistics warehouse, the report showed. As drones took over repetitive inventory counts, employees began focusing on analysing and fixing errors.

On June 26, local time, Business Insider reported that GNC is operating four Corvus Robotics AI drones at its Whitestown distribution center in Indiana, managing reserve inventory in real time.

The drones fly through warehouse aisles, scanning more than 2,000 pallets to automatically verify box counts and storage locations. This allowed GNC to switch to full monthly checks of about 250,000 square feet of reserve inventory from once per quarter.

Bill Monk (빌 몽크), vice president of distribution at GNC, explained: "If the system shows inventory but it has disappeared from the shelf, it leads to a stockout." He said: "If you don't know the inventory location, you can't ship the product to the customer."

The impact also showed up in logistics metrics. GNC said unshipped orders that had previously reached several hundred a day have now fallen to about 98 a day.

The inventory management approach has also changed. In the past, employees spent most of their time walking through the warehouse and counting boxes one by one. Now they focus on checking and correcting inventory mismatches identified by the drones.

Jackie Wu (재키 우), founder and chief executive of Corvus Robotics, explained that the drones fly for about 25 to 30 minutes at a time to inspect designated areas. He said the drones use physical AI that performs autonomous flight and data collection, and that the AI helps customers analyse inventory data and handle errors.

Photos and videos captured by the drones are used not only for inventory checks but also to identify misloaded goods, pallets protruding beyond shelves, and boxes that have fallen to the floor.

Workforce operations have also changed. GNC said it previously employed about 20 staff for inventory counts, but it was not easy to recruit and retain people for repetitive counting work. Most employees have now been reassigned to customer service and inventory management in picking areas, while 1 to 2 people in the reserve inventory area regularly verify the accuracy of drone-collected data.

Monk said about 2 percent of the data the drones judge to be accurate is rechecked by people.

Employees have also responded positively. Tammy Ratcher (태미 라처), who is responsible for inventory management at GNC, said: "In the past, counting boxes all day was most of the job, but now the drones do the counting and we investigate the causes." She said: "Physically demanding work has been greatly reduced."

There was also a case in which AI detected a human error. The drones identified 600 boxes at one location, but the company system listed only 60. It was first suspected to be a drone error, but an investigation found it was a simple data-entry mistake in which an employee left out one '0' when entering inventory.

There are also limitations. The aisles in GNC's reserve inventory area are only about 70 inches wide, leaving little room for drones to fly. If plastic wrap around pallets comes loose, it can catch on the rotors and increase the risk of a crash. As a result, employees fully remove the wrap after work and follow procedures to check aisle conditions before drone operations.

For now, the drones are used only in the reserve inventory area where unopened boxes are stored. It is difficult to directly check remaining quantities inside already opened boxes. Corvus Robotics said, however, that software can estimate remaining quantities in opened boxes to a certain extent.

GNC cited the speed of problem detection as the biggest achievement from introducing drones. Monk said: "If you don't find inventory errors for a quarter or half a year, it takes a long time to resolve them." He said: "AI drones find problems much faster and greatly improve the efficiency of logistics operations."