At AI contact center summit, analysts and vendors shift the measure of success from call speed to issue resolution

The conversation at the summit, covered by theCUBE, centered on how companies can blend human agents and agentic AI. Bob Laliberte of theCUBE Research, speaking with Zeus Kerravala, principal analyst and founder of ZK Research, a division of Kerravala Consulting, said contact centers are an ideal proving ground for agentic AI. "Contact centers have large volumes of interactions. There [are] significant labor costs and direct moments of truth with customers," Laliberte said. "When something fails, the consequences are also pretty highly visible, and a poor AI interaction can increase customer effort, damage trust and ultimately hurt the brand." The two also interviewed executives from TalkDesk Inc., Cisco Systems Inc. and Five9 Inc., among other companies.

Kerravala said the industry has historically judged itself on average handle time and first call resolution, and questioned whether those measures still serve customers. "Those metrics don't … matter as much anymore," he said. "We've had such a focus on average handle time in this industry, but is that shorter call actually a good thing if the issue remains largely unsolved?"

Vinod Muthukrishnan, vice president and general manager of Webex customer experience at Cisco, said the focus needs to remain on relationship management. Cisco's AI Concierge is designed to carry continuous conversational context no matter which "front door" a customer enters at the start of a service interaction. "An AI agent that can fill a slot or fulfill a transaction is not agentic. Agentic is much more of a framework of tools and products and processes," he said, describing a progression from voice agents that answer questions or complete transactions to orchestrating a journey and eventually a relationship.

Ram Rajagopal, head of product for AI at Zoom CX at Zoom Communications Inc., told theCUBE that customer success teams should be measured on "conversation to completion," and that AI agents can help by preserving context across handoffs. "You want to transfer the full context, the reason why they're calling, the collected variables that you need to pass on to a human being so that human being can get on with the job and get it done without having to repeat themselves or look at 10 different places of record just to answer a simple question," he said.

Joe Rittenhouse, co-chief executive officer of Converged Technology Professionals Inc., a Zoom implementation partner, said deployments should begin where they can have the most immediate impact. He pointed to after-hours coverage as a common starting point, describing companies that staff around the clock from the East Coast to the West Coast but leave only a general mailbox overnight, where "the answer is typically, I don't know."

Amit Mathradas, chief executive officer and board member of Five9, described a blended model in which AI takes high-volume, low-stakes calls such as password resets while human agents handle sensitive conversations and high-value customers. That division of labor, he said, could help close the gap between how contact centers measure success and what customers expect from an interaction.