Consumer trust in AI-driven customer service has collapsed, with 68% of respondents expressing no confidence in how businesses deploy generative AI and half reporting they rarely or never achieve successful outcomes through AI-only interactions. The core issue isn't resistance to AI itself—consumers are resigned to its presence—but rather poor execution that creates friction rather than resolution. When chatbots loop endlessly, block escalation paths, or force customers to repeat information, trust erodes even when issues are eventually resolved. This distinction matters: 88% of customers report their problems were solved through AI or hybrid interactions, yet only 22% preferred the company as a result, revealing a critical gap between functional resolution and customer satisfaction.
For CX leaders, this data exposes a strategic misalignment between implementation pressure and customer expectations. Gartner reports 91% of customer service leaders face pressure to deploy AI this year, yet 80% of consumers achieve better outcomes with human-only interactions and 65% actively prefer human-led support. The tension is acute: teams are mandated to implement AI to improve efficiency and self-service metrics, but doing so without robust escalation pathways and reliable resolution capabilities actively damages loyalty. The question becomes whether your current AI deployment—whether through Salesforce's Agentforce, Zendesk's native agents, or third-party solutions—is genuinely reducing customer effort or simply automating frustration at scale.
The path forward requires moving beyond simple chatbots toward what Pega terms "predictable AI agents that consistently get work done." This means designing systems where AI handles what it does reliably whilst maintaining frictionless human handoff for complex issues, rather than treating AI-only interactions as a cost-saving end state. Teams must audit their current implementations against the lived customer experience: are your bots creating dead ends, or are they genuinely accelerating resolution? The data suggests most are doing the former, which means your CSAT improvements may be masking deeper erosion in customer preference and lifetime value.
Your customers aren’t keen on that customer service chatbot you introduced – here's why IT Pro
Your customers aren’t keen on that customer service chatbot you introduced – here's why itpro.com