Study: Consumers Want AI Customer Service That Actually Works
A new survey of 2,000 consumers and 500 enterprise decision-makers reveals a sharp gap between what businesses think AI customer service should do and what customers actually need: resolve their problems.
The research, conducted by Ada and NewtonX between December 2025 and February 2026, found that 59% of consumers prefer instant, 24/7 AI support over waiting for a human agent. But there's a critical condition attached to that preference.
Only 24% of consumers say their most recent AI interaction was fully resolved without human help. The other 76% needed escalation to a person, got only partial resolution, or abandoned the conversation entirely.
What Consumers Actually Value
When asked to rank customer service attributes, consumers prioritized accuracy and problem-solving ability far above empathy. The most common reasons they reported for failed AI experiences were straightforward: the system didn't understand their request (74%), couldn't handle the complexity (56%), or repeated the same unhelpful response (50%).
These aren't failures of tone or personality. They're failures of basic function.
The Business Problem: Misaligned Priorities
Most companies deploying AI in customer service are optimizing for the wrong metrics. Businesses most often cited reduced wait times, lower cost per interaction, and ticket deflection as their primary benefits. Resolution ranked seventh on their list of desired outcomes.
This misalignment stems partly from a measurement problem. Fifty-five percent of businesses lack visibility into how well their AI agents perform. More than half measure AI and human agent interactions together, making it nearly impossible to isolate where the technology is actually failing.
Without that separation, companies can't pinpoint performance gaps, establish a credible return on investment, or identify improvements. They're scaling AI without understanding its effect on the customer experience.
The Skills Gap
Ninety-two percent of organizations plan to increase AI investment in customer experience over the next year. But 36% of customer experience leaders say their internal teams lack the skills and resources to manage, audit, and coach AI agents effectively.
The missing expertise includes agent dialogue design, integration engineering, and knowledge management-specialized skills that most traditional customer service teams don't yet possess.
The full report is available at Ada's research site.
For customer support professionals, the takeaway is direct: your role is shifting. Understanding how to measure and improve AI agent performance-not just deploy it-is becoming essential to the job.
Your membership also unlocks:
