Best Buy Uses AI to Stop Passing Customers Around

A customer seeking help with a malfunctioning laptop might call Best Buy, explain the problem, get sent to a second department and explain it again. If the fix needs a repair, a third agent asks for the same details. Every handoff costs the customer time and costs the retailer another paid conversation.

Best Buy has used artificial intelligence (AI) to cut the number of those handoffs. Transfer rates fell by 1.5% to 2%, and the share of calls resolved without a human rose by more than 50%, the company said in a case study published by Google Cloud. The figures come from Best Buy and its technology vendor and have not been independently verified.

The fix lay in the plumbing. Best Buy connected its customer records to the knowledge bases its agents use, including company policies, troubleshooting articles and product repair data. With all of it in one place, the AI can work out what a caller needs and either fix it or send the call to the right person the first time.

Every Transfer Is a Restart

The problem Best Buy attacked is common. Five9, a call-center software provider, found in a survey of 3,000 consumers and 600 contact-center leaders that 96% of leaders believe their organizations carry context through an AI-to-human handoff. But among consumers, 83% said they still have to repeat themselves at least some of the time after a transfer, CMSWire reported.

Best Buy’s answer is to front-load the diagnosis. The caller with the dead laptop describes the problem in plain words. The assistant—built on Google’s Gemini Flash voice model—works out whether the fault is software or hardware and walks the caller through a repair if one is possible.

The same data reaches the staff. Google Cloud’s Generative Knowledge Assist listens to live calls and puts the relevant troubleshooting document on the agent’s screen as the customer talks. Maureen Spence, Best Buy’s senior director of product, said the goal was to bring the consultative service of its stores to digital and voice channels, and that doing so meant “radically” changing how the company innovates.

Building in Weeks Instead of Months

The second gain is speed. Best Buy moved its customer-care tools to a prompt-driven architecture. CX Agent Studio, the Google Cloud tool its teams use to build voice and chat agents, compresses what the company case study describes as a traditional 18-month development timelines to three to six weeks.

Every voice and chat interaction is logged in BigQuery, and Best Buy analyzes all of it. Teams filter for the products that generate the most support calls and pass the findings to the groups that handle deliveries and product quality. Managers use the same data to tailor training for human agents.

The company has been pointing to results on earnings calls for more than a year. Last May, Best Buy CEO Corie Barry told analysts on the company’s first-quarter call that conversational AI and upgraded phone systems had produced record-low cost per customer contact and record-low transfer rates, alongside record high customer satisfaction. The improvements, she said, were cutting annual call-volume expense.

Best Buy has not framed the system as a replacement for staff. Barry said on the retailer’s 2025 fourth-quarter call that Geek Squad agents would remain available as the AI expands, calling it a dual approach that lets customers “get the support they need on their own terms.”

The same logic runs through the company’s wider AI push. Barry said last year that Best Buy was “leaning into agentic AI not just as a way to improve productivity, but as a way to solve real customer problems,” with store associates and Geek Squad technicians using AI for diagnostics and predictive service.

The strategy will soon have a new owner. Barry steps down as CEO on Oct. 31, and Jason Bonfig, the company’s chief customer, product and fulfillment officer, will succeed her, according to a company announcement in April.