Zendesk on Wednesday introduced a suite of AI-powered tools designed to significantly reduce reliance on human support agents. At its AI Summit, the company unveiled an autonomous AI agent capable of resolving 80% of support issues without human intervention.

The system is complemented by a co-pilot agent to assist human technicians on more complex cases, along with admin-layer, voice-based, and analytics agents aimed at streamlining workflow and decision-making.

Shashi Upadhyay, Zendesk’s President of Product, Engineering, and AI, emphasized the transformative nature of this shift. “The world’s going to move from software built for humans to systems where AI does most of the work,”.

Benchmark testing supports these claims. TAU-bench, which measures tool-calling abilities in AI models, shows top performers like Claude Sonnet 4.5 resolving 85% of support tasks—mirroring the types of issues Zendesk’s AI agents are built to handle.

Zendesk’s move is built on a series of strategic AI acquisitions over the past year, including Hyperarc, Klaus, and Ultimate. The analytics agent in today’s launch leverages Hyperarc technology to provide enhanced insights and data-driven recommendations.

Early trials with existing customers show promising results. “Consumer satisfaction has increased by five to ten points for clients using the AI agents,” Upadhyay noted.

The implications for the customer support industry are significant. Zendesk’s Resolution Platform already handles 4.6 billion tickets annually for nearly 20,000 clients. With AI adoption, the company could reshape workflows for the 2.4 million customer service representatives in the U.S. and millions more globally.

As large language models increasingly handle complex support tasks beyond simple information retrieval, Zendesk’s AI agents may mark a turning point in autonomous customer service.