One of the most appealing aspects of generative artificial intelligence is how it can simplify tasks. From writing drafts to querying databases in natural language, the technology can make formerly complex and time-consuming processes easier. But the simplicity and ease of use on the front end belie a tremendous amount of work information technology teams are doing on the back end to make meaningful use cases.
Graybar, an electrical and communications wholesale distribution business, is in the middle of a multiyear process of modernizing a 20-year-old core system, which started with upgrades to its human resource management tools and is now shifting to enterprise resource planning implementation. The growth of AI presented Graybar leaders with an opportunity to not only upgrade the company’s tech stack but also to think about how to reshape processes to drive new efficiencies and revenue growth. But leaders knew getting the data in order was a must.
“Now we have automation, cognitive AI, all these tools that have so much more power, but that are also making you even more dependent on your data stack,” says David Meyer, chief financial officer at Graybar. “It makes you realize how important it is to have a clean data layer.”1
A big part of Graybar’s modernization effort has been to clean up and standardize data across its core systems. Having consistent data across the enterprise can fuel the kind of cutting-edge applications that generative AI is most known for.
The first areas where Graybar is piloting AI are sales and customer service. It’s adding AI to ordering systems to help surface cross-sell and upsell ideas to sales agents. It’s also developing an AI-based tool that will help agents build quotes for customers. The tool will allow workers to use natural language to query product catalogs, pull together options for customers, and compile the information into a communication for the customer.
“It used to take them hours or days to do that; now it takes minutes,” Meyer says. “AI frees them up to sell or do business development versus spending half a day looking for info and typing up a response to a customer request.”
Building the right data pipelines is one key to establishing effective gen AI use cases; building trust is another. Graybar is rolling out AI in narrowly tailored use cases where tools have access only to specific databases based on what they need to accomplish the assigned task. And in each instance, humans are kept in the loop to help ensure the accuracy of information that comes from AI tools before it reaches customers. Piloting use cases and demonstrating the benefits of AI also help show that the technology is not only effective but also trustworthy. “You have to show people the power that’s there, and what you can do when you get it right,” Meyer says.
For projects that have grown beyond the pilot stage, Meyer has been proactive about tracking their value. He says it’s important to measure these things to ensure leadership continues to support them and that resources are being deployed effectively. He and his team track how much an AI tool has increased margins month to month. For other things that may have less of a directly measurable component, like how much time a tool saved employees, Meyer’s team connects with managers to get frequent feedback. This is all part of the rigorous work it takes to get the most value out of generative AI applications.
“You have to be able to dig in and be disciplined, get good feedback from the field,” Meyer says. “If you’re going to continue to move forward and sell [the tech] to the organization, it’s important to keep track of those value propositions.”