AI-powered return platforms are creating measurable friction in post-purchase customer service, particularly in footwear retail. A Trustpilot analysis of 168,000 shoe reviews from January through May 2026 found that whilst AI represents only 0.3 percent of mentions, those reviews are disproportionately negative—the angriest in the entire category. The core complaint centres on chatbots and automated replies blocking access to human agents during critical moments: 56 percent of dissatisfied customers cited AI-connected customer service as a pain point, with 46 percent specifically frustrated by its impact on returns and refunds. The pattern is consistent across comments: customers describe being trapped in AI loops, unable to reach humans for refund resolution, and receiving only automated responses to urgent delivery and support queries. This represents a fundamental misalignment between where retailers have deployed AI and where customers actually need it.
The strategic implication is stark for CX teams already managing omnichannel support platforms: AI deployment must follow customer intent, not operational convenience. Retailers have inverted the value proposition—using automation to deflect support requests rather than enhance discovery. Trustpilot's data suggests the inverse approach works: AI excels at pre-purchase activities (smart search, virtual try-on, recommendations) where customers actively seek it, whilst post-purchase support demands human escalation paths. For teams running Zendesk, Freshdesk, or similar platforms, this raises a critical question: are your AI implementations reducing friction or creating gatekeeping mechanisms that damage loyalty during high-stakes moments like returns? The data indicates that returns should be treated as loyalty opportunities, not cost-reduction targets, which demands seamless human handoff protocols rather than AI-first deflection.
The broader lesson extends beyond footwear. Customers are signalling a clear operational mandate: deploy AI where discovery and decision-making occur, but ensure frictionless human access where resolution matters. This challenges the assumption that AI chatbots reduce support volume—they may instead be concentrating complaints into a smaller, angrier cohort whilst damaging lifetime value. For CX leaders evaluating AI agent implementations, the question becomes whether your architecture prioritises agent efficiency or customer outcome, and whether your escalation paths are genuinely accessible or merely performative.
Here’s One Big Thing That Frustrates Shoe Buyers About AI WWD