There’s no shortage of vendors promising to transform business communications with AI. What’s harder to find is evidence that it’s working at scale.
Carson Hostetter, EVP and General Manager of AI at RingCentral, sat down with me to talk through where the real traction is, what’s driving customer adoption, and how RingCentral is thinking about the road ahead.
You can watch the full interview here.
RingCentral’s AI Business Is Growing Fast
The starting point for any honest conversation about AI strategy is commercial reality. Ten percent of RingCentral’s revenue now comes from customers who have purchased at least one AI product. Around 15,000 of its 500,000 customers are actively using AI tools, representing roughly $250-260 million of revenue for a $2.7 billion company.
These are customers paying incremental dollars for products they weren’t buying before, which makes the adoption curve more significant than it might first appear.
“You have to imagine there’s a hurdle rate,” Hostetter said.
“Customers are now paying RingCentral more money for something they weren’t buying from us before, and most times weren’t buying anywhere at all.”
Why Context Separates Effective AI from the Rest
Ask Hostetter what makes the difference between AI that genuinely changes a business and AI that simply gets layered on top of existing processes, and his answer is consistent: context.
Not in a technical sense, but in the practical sense of actually understanding what a customer is trying to accomplish before building anything.
“What RingCentral has been really successful at is understanding the context of our customers,” he said.
“A lot of this stuff is pretty basic. We’re in the communications space. Customers want to make sure they get all of their calls answered.”
That grounded approach is reflected in where the company has seen its strongest early results. RingCentral’s AI Receptionist product, AIR, was designed around a straightforward premise: businesses, particularly in regulated industries like healthcare, struggle to staff inbound communications reliably. AIR handles calls, answers questions, and books appointments autonomously, around the clock.
The headline example Hostetter points to is a Colorado mental health clinic now serving 60% more patients since deploying AIR. Staff previously tied to phones are now working directly with patients. Calls that would previously have gone unanswered outside business hours are now handled in real time.
“Unfortunately, mental health isn’t a nine-to-five job,” he said. “Being able to catch those calls and deal with simple questions regardless of when they come in is not only good for the business. It’s good for the patients.”
Business Users, Not IT Teams, Are Deploying This
One of the more telling details in Hostetter’s account is who actually set up AIR at that clinic. Not an IT professional or a developer, but the receptionist.
RingCentral has made accessibility to non-technical users a deliberate design principle across its AI portfolio. The pitch to business customers is that the person who knows the workflows best, not the person who knows the technology best, should be the one configuring the tools.
“You just point it to a website, it scrapes the site, you connect a knowledge base if you have one, give it any guardrails it needs, and you’re off to the races,” Hostetter explained. “We’re not building this for IT or dev shops. That’s a core tenet of our AI, regardless of the product.”
He acknowledged that more powerful products will inevitably bring more complexity, but held the line on the principle. The goal is always to bring configuration back to the people doing the job.
How RingCentral is Measuring AI ROI
On the question of how organisations should measure whether AI is actually delivering value, Hostetter pushed back against the metrics that have historically dominated the contact centre space.
Average handle time, hold time, and abandonment rates were indicators built for a world where the goal was to contain and automate interactions as cheaply as possible. Hostetter’s argument is that AI creates an opportunity to reframe the question entirely.
“The businesses we’re serving aren’t saying ‘great, I can get rid of one person.’ They’re saying ‘great, I can serve twice as many customers. I can talk to twice as many patients.’”
Rather than optimising existing processes, the customers seeing the strongest results are using AI to expand what their business can do.
That framing also shapes how RingCentral is building out the rest of its AI portfolio. AVA, its AI Virtual Assistant, provides real-time support to agents during live calls. ACE, its AI Conversation Expert, analyses interactions after the fact to surface coaching insights and performance patterns. The suite is designed to raise overall business performance, not just reduce operational cost.
RingCentral’s Agentic AI Roadmap
Since this conversation took place, RingCentral has moved quickly to extend its AI capabilities. The company used Enterprise Connect 2026 to announce AIR Pro, a voice-first AI agent platform that brings agentic capabilities into the contact centre, built around a no-code development environment called AIR Pro Studio.
It’s the product Hostetter was pointing toward when he described an “AI Representative” with broader capabilities than the AI Receptionist, able to handle more complex workflows like billing and insurance claims while remaining configurable by business users.
The first vertical supported by AIR Pro is healthcare, consistent with where RingCentral has built its strongest early case studies. Financial services, retail and professional services are expected to follow.
What Organisations Should Take Away
For businesses working through their own AI strategies, Hostetter’s advice is practical rather than prescriptive. Define a real business goal first. Not a cost reduction target or an automation metric, but a genuine growth objective. Then find suppliers you trust enough to treat as partners rather than vendors.
“The breakthroughs we’ve been able to unlock with our customers are built on trust,” he said.
“We earned the right to sit down with those customers and say, let’s explore your business. Here’s what’s new. Let’s look at how we can apply it.”
It’s a position that reflects over 25 years in the communications industry, from early VoIP deployments at Nortel through large-scale contact centre work at Avaya to building RingCentral’s enterprise business from scratch. Hostetter has watched enough technology cycles play out to be sceptical of both the most bullish and the most cautious predictions around AI.
His read is that things will land somewhere in the middle, as they generally do, and that organisations with clear goals and the right partnerships will be best placed to take advantage as the technology matures.
Watch the full interview with Carson Hostetter here.
