75% of AI Customer Service Projects Fail to Prove ROI

Three-quarters of artificial intelligence deployments in customer service fail to demonstrate positive return on investment, creating a troubling disconnect as companies accelerate their spending on the technology.

A Gartner analysis of 432 AI use cases in customer support found that only 25% produce measurable ROI. Another 25% deliver negative returns, while 42% fall into a gray zone where support leaders cannot determine the value generated. Just 11% break even.

Despite these uncertain outcomes, more than three-quarters of customer service leaders plan to increase AI investment in 2026, according to the research released last month.

Why it matters

Customer service leads enterprise AI adoption, with support teams pursuing an average of nearly five AI use cases and allocating roughly 13% of their functional budgets to the technology. Yet as executives tie compensation to AI outcomes — 56% of service leaders expect their incentives linked directly to AI results in 2026 — most cannot prove the technology delivers value. This gap threatens to undermine AI credibility and waste resources on poorly designed implementations.

Top-down mandates miss customer needs

The poor performance stems largely from executive directives that prioritize deploying AI over solving actual customer problems, experts say.

"What we're seeing with the deployment, the backstory here is like everyone is trying to get AI," Antoine Nasr, head of AI at Forethought AI Agents by Zendesk, told Customer Experience Dive. "This is a top-down initiative: We need AI and customer support and customer experience."

Julie Geller, principal research director at Info-Tech Research Group, said too many rollouts begin with board pressure rather than defined business problems. "Too many AI rollouts begin with pressure to demonstrate a credible AI strategy to the board, rather than with a clearly defined business problem," she said.

Flawed assumptions about cost savings

Many organizations expect AI to reduce headcount by handling routine inquiries, but workforce changes tell a different story. About one-quarter of companies report reducing staff while an equal number report growth, according to Gartner. Organizations also need new specialized roles to manage AI systems, offsetting potential savings.

The focus on containment — keeping customers away from human agents — represents another misguided metric. "Containment is also too often mistaken for success," Geller said. "Delaying contact with a human agent is not the same as resolving the customer's problem. The real test is much simpler: did the customer get what they needed, with less effort?"

A separate report by Forethought AI Agents by Zendesk found similar patterns: while 70% of organizations have deployed AI in customer experience, only a small percentage produce meaningful value in outcomes and ROI.

The findings were first reported by Customer Experience Dive.