Challenges in implementing AI in warehouse management

Challenges in implementing AI in warehouse management
The most important challenge in using AI in a warehouse is the proper management of data. As with several other industries, one of the outcomes of using connected devices and AI-based systems is the large amounts of data created. While this data is, without a doubt, crucial to improving operations, no good will come of it if not managed well.

Data management

Naveen Kumar, Director of Customer Service at Aera Technology pointed out that many companies have lots of data, but they don’t know what to do with it, how to handle it, which parts of the data are useful, and which are not, how to use it, etc. This is purely a technology challenge, not a business process challenge, and hence it requires a technological solution.

“In our case, we provide a proprietary solution that can manage the data,” Kumar said. “That’s the first main challenge and it's native to warehouse management where different vendors would use different kinds of systems which may not be easy to integrate with each other. So, making sense of this unclean, scattered, and voluminous data, and learning how the business is operating today and how it can improve in the future. That is the fundamental challenge that we have. This is what we are working towards.”

Connectivity issues

Suresh Yerikireddi, Founder Director & CEO of Inventrax pointed out that lack of proper network connectivity is a major challenge that they face. As warehouses often tend to be outside urban areas where large spaces are available but public infrastructure may not be up to the mark, networks tend to be low-quality.

“We have proven systems that can help optimize operations and resources, but the connectivity plays a key role in their proper functioning,” Yerikireddi said. “Most of the manufacturing plants and distribution centers, which are the data collection centers, are located away from the downtown which often has the data processing center. The connectivity between these two places is critical to the success of AI applications. We cannot bring the manufacturing unit to the city, nor can we take the data center outside the city, so connectivity has to be top quality.”

Convergence of manual and automated operations

Although AI is making a lot of solutions automated, there are still some areas where manual processing is necessary. For best results, there should be a way to ensure that both people and technology are able to work well together.

“For instance, when we are building an item intelligence, we get a manufactured good and receive it at the warehouse,” Yerikireddi said. “Let’s take the case of an electronic warehouse that receives television and refrigerator. When warehouses receive such items, usually employees would take the barcode and stick it on the item or provide an RFID. So, if an operator sticks a TV barcode to a refrigerator and a refrigerator barcode to a TV, the system will not be reading the correct product.

Hence, the convergence of manual and automated systems is a critical stage. You will need to have trained employees who oversee this process and ensure that no human error comes in the way of automation.

Techno-functional specialist

Another major challenge in this sector is finding people skilled in multiple disciplines. This is integral to make sure two different areas of business are able to come together and work hand-in-hand. Unfortunately, the number of people specializing in this area is less at the moment.

“I know you have an operational expert who is good at managing stuff at your distribution point or manufacturing unit,” Yerikireddi said. “I may be a good IT person who knows how to create databases, user interfaces, etc. But there should be a techno-functional expert who knows the operational as well as the IT part of it to understand the overall convergence, the integration of multiple data sources, and multiple data systems.”


Product Adopted:
Other


Share to:
Comments ( 0 )
security 50