Tractor Supply enlists AI to deliver ‘legendary’ customer service

Tractor Supply Co. prides itself in delivering “legendary” customer service, and it has turned to artificial intelligence to assist with that goal.

The rural lifestyle retailer, with more than 2,200 stores across the US, caters to the needs of rural residents with a wide variety of products on its shelves. Shoppers can find overalls, live chickens, gun safes, ice fishing augers, ferret food, throw blankets, a farmhouse-shaped egg holder, and welding machines as they walk the aisles of most Tractor Supply stores.

And while the company prides itself on the wide knowledge of its store employees, no one can be an expert on every product a store sells. A new support tool, Hey GURA, combines generative AI with headsets to provide store employees with all kinds of information on the fly. With Hey GURA, a store employee can immediately call up product specs, such as the amount of hot air the PelPro Pellet Stove can move per minute, without seeking out a computer terminal.

Hey GURA, similar in some ways to Apple’s Siri and Amazon’s Alexa, can also offer product recommendations that employees can pass on to customers, and it can provide employees with the locations of products in the store, as well as HR and company information.

With information about products and availability constantly changing, Tractor Supply sees Hey GURA as a “knowledge base and a training platform,” says Rob Mills, chief technology, digital commerce, and strategy officer at Tractor Supply.

Hey GURA is “really leaning in on that legendary service but also building that authority and trust that we already have with the customer,” he says. “It makes the team member much more efficient.”

Explaining life out here

The Hey GURA assistant includes a wide-ranging “life out here” knowledge base, echoing Tractor Supply’s corporate brand message. The knowledge base offers information about land, animals, and other rural topics customers may ask about, even when the request isn’t about a product in the store, says Glenn Allison, vice president of IT enterprise architecture and IT solutions.

Hey GURA includes a history of past questions and answers it has tackled, and it can learn to provide customized answers for customer queries in different parts of the country, Mills says. Customers may have questions about plants to grow in a certain climate zone, for example.

Rob Mills / Tractor Supply Co.

“It’s about having the right information at the right time at all times for that customer that leads back to that legendary service,” Mills says.

In addition, Hey GURA has become part of Tractor Supply’s new employee training program, Mills adds. “One of our differentiators is we hire team members from the community,” he says. “We’re always hiring and growing and looking at new team members. With this tool, you’re able to get a team member very efficient very quickly.”

Identifying customers who need help

Meanwhile, a second AI-driven project, called Computer Vision or Tractor Vision, uses AI and in-store cameras to alert store employees not only when lines are building at check-out registers but also when browsing customers look like they need help, with the goal of eliminating the need for customers to search for an employee. If a customer is browsing alone in the garden center, the Computer Vision AI can alert an employee with gardening expertise to meet the customer there.

“It’s about how we make sure we match up the customer team or the team member’s skill with the customer when they need it,” Mills says.

Computer Vision also gives insights about customer traffic in stores, including metrics on conversion rate and sales numbers, thus helping the company decide what products to stock, how to display them, and how to arrange products across the store, Allison adds.

Together, the two projects earned Tractor Supply a 2024 CIO 100 Award for IT leadership and innovation.

Move fast, be nimble

The Tractor Supply tech team moved quickly with both projects, completing the Computer Vision project in six to eight months and taking Hey GURA from concept to a prototype launching in 200 stores in about three months, Mills says.

“Our philosophy in general is, ‘Let’s be nimble,’” Mills says. “The tech team is extremely agile.”

Tractor Supply already had several foundational technologies in place when the projects started, helping to speed development, Allison says. All Tractor Supply stores already had SD-WAN networks and 5G broadband service available, and security cameras were IP enabled, enabling them to link into the Computer Vision capabilities. 

Both projects were rolled out to pilot stores in late 2023, and Hey GURA was fully deployed to all 2,200-plus locations in 49 states earlier this year. Store employees can access Hey GURA through either AI-enabled communicators and a headset or an app on employee mobile devices.

Computer Vision is currently deployed to more than 100 stores and is being rolled out to all stores, with a goal of reaching 700 stores by the end of the year. As part of the rollout, each store is upgraded with an edge computing solution that leverages a container solution for point-of-sale devices, the management of security cameras, AI voice communications, and AI models with computer vision.

Tractor Supply also uses a secure cloud platform, along with edge computing, for AI engineering projects and the development of AI-based applications. The tech team uses AI machine learning operations (AI MLOps) and AI large language model operations (LLMOps) practices.

Innovating with AI

Tractor Supply’s use of AI tied to other technologies appears to be innovative in the retail sector, says Sucharita Kodali, a retail industry analyst and vice president with Forrester Research. While some grocery stores have experimented with AI to cut lines at checkout registers, Kodali hasn’t seen another project pairing AI with headsets or mobile phones to assist with customer questions.

The use of AI in the retail sector is in the experimental stage, and many retailers have gotten only as far as using AI-powered chatbots on websites, Kodali adds. It’s still difficult, although not impossible, to calculate the return on investment for AI use in retail, she says.

“It’s always been an expensive technology,” she adds. “Everybody would love to put their associates where they’re most wanted. There was never a really efficient way to do it.”

It will be interesting to follow Tractor Supply’s projects and see how the company calculates value from Hey GURA and Computer Vision, Kodali says. The two projects seem much more useful than other retail experiments, such as AI-generated billboards, she suggests.

“I applaud them for trying,” she adds. “To their credit, they’re actually focusing on real customer experience opportunities, which is more than I can say for a lot of random AI experiments.”