AI in CX: Only 1 in 10 Leaders Report Significant ROI

What happened

A new industry survey has found a striking disconnect between AI spending ambitions and measurable results in customer experience: just one in ten CX leaders report that AI has delivered significant impact on their operations to date. The findings, reported by CustomerThink, reveal that the majority of organisations are pressing ahead with investment despite limited evidence of returns.

According to the survey, 57% of CX leaders say AI has had limited or no impact on their CX operations over the past year — yet 62% expect their AI investment to increase over the next twelve months. The pattern suggests that optimism, competitive pressure, or fear of falling behind is driving spend decisions more than demonstrated performance.

Why it matters

For customer experience practitioners, this gap is a behavioural economics story as much as a technology one. The continued willingness to increase investment in the face of poor returns echoes the sunk-cost fallacy and status quo bias — organisations may be doubling down on AI because abandoning it feels riskier than persisting, regardless of what the evidence shows. In service design terms, deploying AI without a clear measurement framework is the equivalent of redesigning a customer journey without ever speaking to a customer: activity is mistaken for progress.

The findings also carry a warning for CX leaders under board pressure to "do something with AI." When investment outpaces understanding, the risk is not just wasted budget — it is degraded customer experience. Poorly implemented AI in service environments can erode trust, increase friction and damage the emotional quality of interactions at precisely the moments that matter most to loyalty.

By the numbers

• 1 in 10 CX leaders report significant impact from AI on their operations to date.
• 57% of CX leaders say AI has delivered limited or no impact on CX operations over the past year.
• 62% of CX leaders expect their organisation's AI investment to increase over the next year.

The Renascence take

The headline numbers will be read as a cautionary tale about AI hype — but the more important signal is structural. Most organisations are measuring AI impact the wrong way, or not at all, which means the 57% reporting limited results may actually be underestimating failure while the 10% claiming success may be overcounting it.

The real problem is not that AI underdelivers — it is that CX teams adopted AI before defining what "impact" means for their customers. Significant impact cannot be self-reported; it must be traceable to a specific customer outcome: reduced effort, faster resolution, higher trust, or demonstrable loyalty uplift. Until organisations instrument their AI deployments with behavioural outcome metrics rather than operational proxies, the investment-to-ROI gap will persist no matter how much spend increases. The contrarian move right now is not to invest more — it is to measure better, even if that temporarily slows the rollout.