Zoom has unveiled a wave of ZoomCX innovations that challenge the metrics contact centers have relied on for decades.
Announced at Enterprise Connect 2026 under the banner of what it is calling the “resolution economy,” the vendor is aiming to introduce something it believes better reflects whether a customer interaction was actually successful.
Average handle time, containment, and deflection rates have long been the CX stalwarts. Yet, in many instances, despite these metrics looking healthy on a dashboard, the actual level of customer experience can still be significantly below par.
Zoom's argument is that those numbers measure efficiency, not outcomes – and that for too long the industry has confused the two.
Chris Morrissey, General Manager of Zoom CX, framed the stakes plainly in the official announcement:
“The difference between losing a customer and earning their loyalty comes from interactions that build real relationships, and don't just handle tasks.
“Our focus is on helping organizations build those connected relationships that reduce effort, solve problems fully, and turn every interaction into long-term value for their customers.”
Michelle Couture, Head of Customer Experience Product Marketing at Zoom, expanded on the metrics shift during the briefing.
“Historically, contact center was really optimized for speed, using metrics such as average handle time, hold time, containment, deflection,” she said.
“Those metrics are still really important, but when you're thinking about resolution modeling, you need to start thinking about a broader set of outcome-based indicators.”
For Couture, that means tracking first contact resolution, repeat contact rate, time to resolution across systems, and customer effort score.
“Resolution is not just measured on how quickly a ticket closes, or if it was contained,” she added, “but really about whether a customer's need was fully addressed and the context carried forward without friction.”
The new launches are designed to give contact center teams the tools to actually deliver on that standard.
Zoom Virtual Agent 3.0
The headline product is Zoom Virtual Agent 3.0, which launched last week ahead of the event.
Built on a multimodal large language model foundation, ZVA 3.0 can process voice, chat, images, and structured inputs simultaneously.
In practice, that means it can interpret visual information – such as serial numbers, documents, and photos – removing one of the more persistent friction points in self-service channels.
Where ZVA 3.0 departs more significantly from its predecessors is in how it operates. As Couture put it during the briefing, “it doesn't just respond, it reasons and executes.”
ZVA 3.0 can orchestrate actions across systems, trigger real-time tasks in third-party applications, and provide full journey transparency so that administrators can trace every AI decision for accuracy and compliance purposes.
A built-in knowledge base suggestion feature captures human-validated resolutions and applies them to future cases – giving contact center leaders a way to scale consistent service quality without losing oversight.
Zoom also points to its own internal deployment of ZVA as a proof point: CSAT increased by 25 points after implementation.
Underpinning ZVA and the broader CX platform is what Zoom calls Federated AI, a distributed reasoning architecture built to support connected customer intelligence.
In Tau²-bench evaluation testing, Zoom's self-improving agent achieved 92.8% first-pass accuracy on retail-domain tasks and outperformed leading models – a number Zoom is leaning on as it makes the case for agentic AI in enterprise contact centers.