Quick Overview

  • According to a survey released last week by Laivly, a contact center AI and automation platform, two-thirds of customer experience (CX) leaders say their recentAI projects have been successful, but many are hitting major obstacles.
  • Slightly more than half (53%) of AI projects have exceeded budget, and 43% are currently delayed or stalled.
  • Some AI projects are also hurting revenue: 28% of leaders attribute revenue loss to AI's inability to handle customer complexity; another 20% say revenue loss is occurring but cannot quantify the extent of the damage.

Deep Insights

There is a gap between how CX leaders perceive their AI projects and actual results—Ian Elliot, director analyst in Gartner's Customer Service and Support team, attributes this paradox to the pressure service leaders face to adopt AI.

"From our observations, the gap in leadership's perception of AI success stems from a combination of immense market pressure, misaligned personal incentives, reliance on flawed performance metrics, and insufficient accuracy of AI tools," Elliot told CX Dive in an email.

Elliot noted that because investors actively reward companies that claim AI success and cost savings, CEOs are pressuring service leaders to deploy AI and meet unrealistic expectations.

Personal financial interests exacerbate this situation. Gartner found that in 2026, more than half (56%) of service leaders will have incentives directly tied to AI outcomes.

"In a rush to demonstrate expected cost savings, some organizations have even prematurely cut staff, merely to free up funds to finance their AI ambitions, rather than reducing headcount because AI deployment has truly succeeded," Elliot said. "This creates a surface narrative of cost-cutting success at the executive level, but it's not as evident when you look at the hard data."

AI tools can be useful, but too many areoverhyped and fail to meet expectations。

A Sinch survey in May found that three-quarters of businesses havepulled back AI deployments. Key reasons for the pullback include: customer data exposure, hallucination or brand risk, and inability to diagnose where problems lie.

Respondents in the Laivly survey also highlighted such issues. One-third of leaders said AI tools pose compliance and tone risks, and 36% said agents find the tools difficult to use due to a lack of cross-interaction context.

"While leaders report success based on usage and headcount reduction, the reality is that the technology often fails to support employees," Elliot said. "Fewer than half of agents believe their current systems are reliable, and only 56% trust the accuracy of information these systems provide."

This leads agents to repeatedly verify information provided by AI, thereby slowing down the efficiency gains leaders assume AI will bring.