If you’re concerned about how AI will disrupt the job market, you’re not alone. A recent study by Gartner showed that about three in four employees believed AI would cause a net decrease in jobs. As indicated by that same study, however, where businesses had implemented AI, only one in eight thought the same. As it stands, AI augments rather than replaces, with computer vision (CV AI) being able to aid operational teams in many of their day-to-day tasks.
While the technology is helping to decrease human error in inventory and improving the customer experience in brick and mortar stores, the expertise and experience of operational personnel will still be required to ensure its abilities are applied in a valuable manner.
Here are 3 reasons why AI will augment but not replace logistics teams:
1. Computer vision can help detect safety hazards, but humans must first teach it what safety hazards look like.
From slips and falls to equipment malfunctioning, warehouses can be dangerous environments. For instance, nearly 100,000 workers are injured in forklift accidents, with almost 34% of those workers suffering injuries. With computer vision, however, these hazards can be detected before injuries can occur. This technology can quickly analyze security footage, alerting the relevant persons about hazardous site conditions and where each worker is located. 
The vast majority of fatal forklift accidents are caused by forklifts tipping over and crushing workers, with everything from driver error to debris causing these mishaps. CV AI technologies like object detection and object tracking can help managers to monitor warehouses for issues such as debris in the path of the forklift or damage to the forklift itself to be detected before accidents can occur.
The above said, on its own, computer vision won’t be able to recognize what these hazards are. The expertise of operational personnel is required to train CV AI properly so that it can know when safety standards are breached.
2. Computer vision can make inventory management more efficient, but humans are still the problem solvers.
Thanks to barcode scanners, the days of relying on employees to physically count your inventory are long behind us. CV AI, however, takes this advancement a step further allowing us to teach models to recognize barcodes. By integrating such a model with a barcode scanner, multiple barcodes could be read at once, decreasing time spent scanning each one.
CV AI can also be taught to recognize products by their appearance, enabling you to create consistent tags throughout your inventory, so you can easily search for items in your stock management platform. CV AI models can also learn that products may have different names (“chips” versus “fries,” “soda” versus “pop”), so you or your team can find the desired products, no matter what term is used.
Many businesses across industries are already using AI for quality control, helping them to identify defective products before they leave the warehouse or get into the customers' hands. The fact remains, however, CV AI can’t learn to recognize these things on its own. As I said above, the expertise of your personnel will be required to ensure the model is trained adequately to recognize, for instance, produce that doesn’t meet an aesthetic standard. What’s more, while CV AI can alert you about problems, like excess stock, operational teams will have to come up with dynamic solutions to those issues, like sales.
3. Computer vision can gather all the data needed to make stores more efficient, but humans are required to interpret these results.
As operations managers look to make their stores more efficient, CV AI allows you to gain concrete data to support this using the technology you already have. Powered by CV AI, both live and recorded security camera footage can be leveraged to monitor and measure foot traffic patterns and identify heavy traffic areas, so an operations manager can determine whether an adjustment to the store’s layout is required.
By using facial recognition to identify staff, this same footage can be used to recognize when an unauthorized person, like a customer, enters a restricted area to prevent problems like theft. It can also augment your employees as they seek to meet customer service standards. Facial recognition can be used to analyze a customer’s face and gauge their satisfaction. If dissatisfaction is detected, an employee is notified, allowing them to swoop in and repair or improve the experience. When combined with the foot traffic and traffic flow data, this information may also help operations managers see whether the store’s operations or layout contributed to the customer’s ill-feeling.
AI has already proven itself to be a valuable tool for data analysis as it allows for the consolidation of data from different sources, such as visual data from various stores with that of footage from the relevant warehouses. CV AI allows for this data to be quickly filtered for the relevant data, providing managers with an in-depth view that may help them to identify patterns or opportunities that would otherwise go undetected.
It is the expertise of operational teams that will determine whether all this data is leveraged to your advantage. While CV AI can take over cumbersome rote tasks, like data cleansing, in its current form, it cannot turn what it learns into something meaningful on its own. That task is still best left to the humans who use it.
While CV AI’s potential to improve operations cannot be downplayed, operational teams have no reason to fear it. Much like the abacus replaced counting on fingers and the calculator displaced the abacus, CV AI will only displace the tools we’ve created to make our lives easier. It is here so we can direct our uniquely human intelligence towards making the world operate efficiently and to our benefit.
