A Gartner analysis of 432 AI customer service use cases reveals a stark disconnect between investment appetite and demonstrated value: only 25% produce measurable ROI, whilst another 25% actively destroy value, and 42% operate in a measurement vacuum where leaders cannot articulate returns at all. Just 11% break even. This creates a peculiar paradox in the CX technology market—despite these findings, over three-quarters of support leaders plan to increase AI spending in 2026, suggesting either institutional momentum overriding data, or a fundamental misalignment between how organisations measure success and what their AI implementations actually deliver.
The implications for CX teams are twofold. First, this data should prompt immediate scrutiny of existing deployments: if your organisation falls into the 67% experiencing unclear or negative returns, the problem likely isn't the technology itself but rather use case selection, implementation rigour, or measurement frameworks. Teams running mature platforms like Zendesk or Salesforce Agentforce should audit whether they're applying AI to high-volume, repetitive interactions where ROI is demonstrable, or spreading investments across marginal use cases that consume resources without proportional impact. Second, the continued spending despite poor returns suggests that many leaders lack the diagnostic tools to distinguish between genuinely underperforming implementations and those simply poorly measured—a gap that creates risk for both vendors and enterprises as capital flows toward solutions without clear accountability.
The broader question this raises is whether the CX industry has developed adequate measurement discipline for AI initiatives. When 42% of use cases have "unclear ROI," that's not a data problem—it's a governance problem. Teams should demand that vendors and internal stakeholders establish baseline metrics before deployment, not retrofit measurement afterwards. The organisations capturing value from their AI investments are likely those treating implementation as a controlled experiment with defined success criteria, not as a technology adoption exercise.
Customer service teams are increasing investments in the technology even as economic value remains elusive, Gartner finds.