Customer Experience13 July 2026

AI agents in customer support: the first production benchmarks are in

Autonomous resolution of 40-60% of tier-1 tickets, stable or rising satisfaction: AI agent deployments in customer support reach maturity. Success conditions and observed pitfalls.

The first benchmarks of at-scale AI agent deployments in customer support are now public: 40 to 60% of tier-1 tickets resolved without human intervention, first-response times divided by ten, customer satisfaction stable or slightly up when the setup is well designed.

The discriminating factor observed by Cardan-AI: knowledge base quality. Performing agents rely on up-to-date, structured, arbitrated product documentation; failures almost always come from obsolete or contradictory knowledge the agent restates with confidence.

Second condition: frictionless escalation routing. The setups that gain satisfaction are those where the agent hands over to a human as soon as confidence drops, with full conversation context. Those that trap the customer in an automated loop destroy more value than they create.

The full analysis is reserved for Business members

Sector impact, operational recommendations and key figures.

Become a Business member

Analysis by

Cardan-AI Intelligence

Our research and analysis unit, dedicated to applied AI for business, industry and regulatory compliance.

Sources

Cardan-AI — Support agent opportunity study: request a call

Let's talk about your next competitive edge

A 30-minute conversation to identify your most profitable AI use cases.

Book a meeting