Watch: How AI Addresses Inventory Challenges

Artificial intelligence can help eliminate unneeded inventory, says Sankalp Arora, chief executive officer and co-founder of Gather AI.

It’s not news that e-commerce ballooned during the pandemic, but it was a revelation for some that each piece of inventory gained importance because every missed SKU was a failed order and a lost revenue opportunity. “With the shift to each individual SKU being shipped out to people,” Arora says, “inventory monitoring and the need for inventory accuracy became compelling.” 

As an example, he says, if scanning fails to locate an order for a pair of jeans, an inventory quality assurance team has to investigate the mis-pick. In the meantime, a truck has to leave because same-day or two-day orders have to be met. 

“You just missed the order,” Arora says. “And not only that. On that same order where you might have spent around a dollar, now you're spending $5 in fulfilling that order. How do you overcome that? If you don't have that precision, you carry shadow inventory, which is way more than you need just to cover up for those mistakes. That ends up sitting on your balance sheet. No CFO wants that.”

Arora says there are three areas where AI can improve matters. One is in physically moving goods within the warehouse through autonomous forklifts and item picking. Second is in being able to better predict how much inventory is needed. “Even Amazon recently said that their main initiative for the next year is to get better prediction so that they can get on the right inventory levels,” he says. Lastly, there’s a need to feed in the right data on what's sitting in the warehouse.

Computer vision tracking of inventory can give real-time updates on how things are moving. “You know exactly what's sitting where in your warehouse,” Arora says.