UponAI Wants to Fix Voice AI by Starting with the Phone System, Not the App

Here's a question most AI companies never ask: does the AI actually understand how phones work?

Sounds basic. But according to Bill McClain and Jody Gonzalez of UponAI, a lot of voice AI platforms skip that step entirely. They build something clever, bolt it onto a phone line, and hope for the best.

UponAI took the opposite approach. They started with telephony first, AI second.

I like that order. It reminds me of something I've learned moving around the world for the better part of a decade: the smartest solution usually works with what's already there, not against it.

Built From the Partner's Side of the Table

McClain and Jody Gonzalez sat down for a recent Technology Reseller News podcast to explain how UponAI is giving managed service providers a way into voice AI without blowing up a customer's existing phone system.

"We are our partners," they said. "We've seen what they see, and we built the platform around those real-world problems."

That's not a tagline. It shows up in the architecture. UponAI connects to a customer's existing PBX through SIP extensions, so the AI agent operates inside whatever phone environment is already running. No rip-and-replace. No forcing a client onto a new platform just to get AI features.

For MSPs working with customers who have mixed, legacy or constantly shifting communications setups—which, let's be honest, is most businesses—that matters a lot.

Gonzalez points out that many voice AI tools were designed by people who understood AI but not telephony. UponAI flipped that. The AI needs to understand extensions, queues, transfers, call flows. The phone system comes first.

A Toolbox, Not a Fixed Product

UponAI describes itself less like a finished product and more like a toolbox MSPs can use to build their own agents. It connects to CRM, billing and accounting systems through APIs, MCP, webhooks, and low-code or no-code tools.

The whole thing is white-label too, so partners can put their own name on it and set their own pricing.

Speaking of pricing, here's a detail that caught my attention: per-second billing. Instead of rounding every short call up to the next full minute, UponAI charges for the actual seconds used. McClain says at higher call volumes, that difference adds up fast. Small detail, real money.

Outcomes Over Gimmicks

What strikes me most is how much emphasis UponAI puts on actually solving the caller's problem, not just sounding smart.

The AI agents can tell if someone's already on another call, handle smarter transfers, and circle back to the caller if the person they need isn't available—rather than just shoving them into voicemail and calling it done.

"It's about outcomes," McClain says. "If the customer's problem isn't solved, what good is it?"

That's a question worth asking about a lot of technology, not just phone systems.

Real Use Cases, Right Now

This isn't theoretical. One upcoming deployment is a medical office where patients will be able to verify their identity and settle outstanding balances through an AI agent.

Another is an operation at JFK Airport handling international calls related to pets—imagine the range of languages and accents involved there. UponAI's agents can automatically switch between 10 languages, with more available through separate agents. Gonzalez also mentioned pronunciation controls, which matter when you're dealing with names, cities, or technical terms that most AI tools mangle.

I've spent years navigating language barriers in airports, customs lines, and phone calls across a dozen countries. Getting pronunciation right isn't a nice-to-have. It's the difference between being understood and being hung up on.

Getting Started

For MSPs curious about jumping into voice AI, UponAI offers a four-week onboarding process to help partners build their first agent and learn the platform well enough to manage future builds themselves.

Four weeks to go from curious to capable. That's a timeline worth paying attention to if you're an MSP wondering whether voice AI is worth exploring—or if you're just someone who likes watching how new tools get built around real problems instead of imagined ones.

Either way, it's one more sign that the tools shaping how we communicate keep getting more flexible, not less. And that's always worth a closer look.

Source: Telecom Reseller / Technology Reseller News

If you missed it: Turn Your Next Conversation Into Your Best Marketing Asset