RingCentral’s latest earnings call hinted at a shift many CX leaders already feel in their bones. The real value in AI is moving from “after the call” to “during the call.”
On the same day as its Q4 2025 earnings update, the vendor also announced a collaboration with OpenAI to deliver ‘advance enterprise-grade voice AI,’ referencing models ‘like GPT-5.2,’ with an emphasis on real-time performance.
That product news also sharpens one of the most telling metrics from the earnings call: RingCentral’s AI Conversation Expert (ACE) customer count now exceeds 4,800, up 144% year-over-year. Together, the two updates point to an agent experience (AX) story that is starting to harden into an operational reality.
OpenAI’s Role: From Insight To In-Call Action
RingCentral’s OpenAI integration announcement pushes the conversation forward from analysis to action.
The company says it is integrating OpenAI to ‘advance enterprise-grade voice AI,’ and it calls out models ‘like GPT-5.2,’ with an emphasis on real-time performance, as Kira Makagon, President and COO at RingCentral highlighted: “OpenAI enables us to turn powerful technology into tangible business value, from AI that answers customer calls to AI that assists every employee.”
Giancarlo “GC” Lionetti, Chief Commercial Officer at OpenAI, also leaned into the live-workflow angle:
“RingCentral shows how advanced AI moves beyond insight and into action. By working together, we are bringing intelligence directly into live voice conversations, helping enterprises move faster, serve customers better, and act with confidence.”
For AX, this matters because it accelerates a specific operational pattern: real-time agent assistance that behaves like a coach in the ear.
In practice, that can mean next-best-action prompts, compliance reminders, better discovery questions, and objection handling that arrives while the customer is still on the line.
RingCentral’s own product positioning supports that direction. The OpenAI announcement describes natural, context-aware interactions and ties them to RingCentral’s ‘low-latency voice infrastructure,’ which is a requirement if guidance is going to be useful in the moment.
The New CX Battleground Is The Live Call
Contact centers have spent years optimizing what happens after an interaction. Teams built QA scorecards, coaching cadences, and post-call analytics, and they also chased better summaries to reduce admin time.
But the friction point has stayed stubbornly consistent: the call itself is where outcomes are won or lost, and where agents need support most.
RingCentral’s AI portfolio framing makes this ‘before, during, after’ model explicit.
Makagon described how AIR, AVA, and ACE align to different moments of the conversation lifecycle, and she placed ACE firmly in the improvement loop after the call, stating:
“After the call, our AI Conversation Expert, or ACE, closes the loop, analyzing every recorded interaction for insights that improve coaching, quality and performance across the organization.”
That line sounds familiar, but the growth figure attached to it is the part that should make CX leaders pause. A 144% year-over-year increase suggests that more organizations are treating conversation intelligence as foundational to performance management, and also scaling it across sites, queues, and roles.
Why RingCentral’s 144% Metric Looks Like A Market Signal
A single vendor metric is not a market study, and RingCentral’s ACE number is a customer count growth figure, not a disclosed usage volume metric.
Still, the direction is hard to ignore, and it maps to what many operations teams are prioritizing: better consistency, faster ramp, and fewer escalations, even as channels, products, and policies keep changing.
RingCentral also paired the ACE story with operational outcomes from customer examples, including Patient Connect, which Makagon said uses ACE quality management to reduce escalations:
“They also use ACE quality management to replace time-consuming spot checks of call recordings, reducing escalations by 40%.”
That is a classic QA and coaching use case, and it also hints at a broader change in how quality gets managed. Instead of sampling a small fraction of calls, teams increasingly want coverage across every recorded interaction, and they want the findings to flow into coaching quickly.
And while ACE is positioned as post-call, RingCentral’s ‘before, during, after’ portfolio pitch sets up a bigger move: the loop tightens until coaching becomes real-time guidance, and post-call review becomes the audit trail, not the primary coaching moment.
What ‘Real-Time Guidance’ Looks Like In A Modern Contact Center
If we zoom out, ‘real-time guidance’ is becoming less of a premium feature and more of an operational requirement, especially in environments where:
Agents handle complex calls, and policy changes weekly
New hires ramp constantly, and attrition remains a reality
Compliance requirements are strict, and mistakes are expensive
Leaders need measurable, repeatable behavior change
RingCentral’s earnings call also reinforced why latency matters in this transition. In Q&A, Makagon described an approach that optimizes for accuracy, latency, and fit-for-purpose model selection, and she emphasized that the platform is “model agnostic.”
That matters to buyers because real-time assist fails fast when latency is high, and it also fails when accuracy is inconsistent. If RingCentral can orchestrate models based on the task, and keep performance stable, then live coaching becomes easier to operationalize at scale.
The Agent Experience Question CX Leaders Should Ask Next
The headline promise of generative AI in contact centers has often been about reducing after-call work. That is valuable, and it is also increasingly table stakes.
The more strategic question is this: Are we improving agent performance at the moment it matters, and can we prove it?
RingCentral’s ACE growth figure suggests more teams are investing in coaching and QA infrastructure, and its OpenAI integration suggests it wants that infrastructure to reach into the live call itself.
If that trajectory holds, agent experience will stop being a soft metric tied to sentiment surveys, and it will become an operational discipline tied to real-time enablement, measurable behaviors, and consistent outcomes.
And if that happens, the best contact center AI will not be the one that writes the cleanest summary. It will be the one that helps your agents win the call while it is still happening.
Join the conversation: Join our LinkedIn community (40,000+ members): https://www.linkedin.com/groups/1951190/
Get the weekly rundown: Subscribe to our newsletter: http://cxtoday.com/sign-up
