Customer Loses $1,700 Ebike to AI Chatbot Maze at FedEx

When a $2,000 ebike disappeared after FedEx marked it delivered and signed for by unknown initials, an Atlanta customer expected the resolution process would be straightforward. Instead, what followed was a three-month odyssey through AI-powered customer service systems that couldn't—or wouldn't—connect him to anyone who could help.

The bike never arrived. FedEx's chatbots repeatedly blocked requests to speak with humans. The local police department routed calls through automated systems that never followed up. The bicycle retailer offered only shipping reimbursement—one-tenth the purchase price. Credit card and bank disputes hit dead ends when representatives finally surfaced, only to say the problem belonged to FedEx. The customer remains out $1,700.

The automation wall

This case illustrates a broader shift in corporate customer service. According to an April survey of customer service leaders, 31 percent have already reduced or plan to reduce headcount due to AI adoption. Verizon's CEO recently told Bloomberg that AI will likely replace a "large percentage" of customer service work, calling it one of the sectors most exposed to the technology.

The result for consumers is what researchers call "sludge"—friction intentionally or unintentionally designed to discourage people from seeking resolution. Ryan Hamilton, a marketing professor at Emory University, notes that while sludge predated AI, the technology has "ramped up the dystopian nature of it."

Consumer sentiment reflects the disconnect. A May report covering the US, UK, and Canada found 59 percent of shoppers frustrated with AI customer service agents, while 85 percent said they'd prefer speaking with a real person.

Why it matters

Companies are making massive AI investments—global spending on AI tools is expected to surge this year—but many are deploying chatbots before the technology can handle complex, high-stakes problems. Ravi Dhar, director of Yale's Center for Customer Insights, describes a sunk-cost fallacy at work: executives face pressure from investors to show AI strategy and returns, potentially leaving them committed to implementations that aren't working. Meanwhile, customers absorb the cost of undertrained systems that can't escalate appropriately or access the right data to solve problems.

Hamilton warns that companies may be operating on "optimism," assuming AI will eventually catch up. But in industries where service quality matters, the reputational damage from poor experiences could outweigh efficiency gains. The risk is that AI homogenizes customer service across sectors, eliminating it as a competitive differentiator just as consumers need it most.

The human gap

When reached for comment, FedEx acknowledged the tension, stating that "complex situations require human care and deeper support" and that the company is "continuously refining our processes." The Atlanta Police Department confirmed that returning missed calls routes callers back to the chatbot-run non-emergency line.

For the customer still missing his ebike, the lesson is clear: current AI customer service systems excel at deflection but fail at resolution when things go wrong.

These details were first reported by WIRED.