Inside Tavus: How Quinn Favret Is Building the Operating System for Human-AI Interaction

In this episode, Stacey talks with Quinn Favret, COO and co-founder of Tavus, a platform for creating lifelike digital humans that enable natural, real-time conversations between people and AI.

Quinn shares how his early passion for building evolved into a mission to make machines communicate more like humans. He explains what Tavus actually does, how it works, and why it’s different from tools like chatbots or AI SDRs.

They explore real-world use cases across healthcare, education, and go-to-market, including AI therapists, tutors, digital sales reps, and customer success agents. Quinn also offers insights on how companies of all sizes are using the Tavus platform to scale human-like interactions in ways that weren’t possible before.

Whether you're a marketer, founder, or curious about where AI is heading, this conversation breaks down what it means to build AI that feels human; and how it can amplify human creativity, not replace it.

Transcript

Stacey Epstein (00:33)
Today on the show, we have Quinn Favret, the COO and co-founder of Tavus. Welcome to the show, Quinn.

Quinn Favret (00:42)
Hey Stacey, great to see you. I’m happy to be here—amped, actually.

Stacey Epstein (00:46)
Awesome. Happy to have you. If anyone's watching us on YouTube, you'll see I'm in a different location today, and the Wi-Fi isn't great. So I might be coming and going, but I think Quinn has a strong connection—and that’s what matters. So let’s get going. Quinn, tell us about you. What’s your journey? What did you do before Tavus, and how did you get here?

Quinn Favret (01:03)
Great question. It starts a long time ago. My grandparents immigrated from Armenia to Michigan during the genocide, so I was born and raised in Michigan. My dad didn’t go to college, and my mom was an accountant. Growing up, I was always taught the value of hard work and making things happen for yourself.

But unlike a lot of founders who grew up coding, I was building in different ways—robotics in the garage, treehouses in the woods. I’ve always loved piecing things together. For a while, I thought politics was the way to make change, so that’s where I started.

Stacey Epstein (01:48)
Wow.

Quinn Favret (02:01)
Yeah—turns out, not so much. I started working on Capitol Hill early on, doing policy work. And pretty quickly, I realized the idealistic view I had of politics didn’t match reality. So then I went to college and rediscovered my love for building—this time through technology.

That led to Startup Idea Number One. It wasn’t intentional; I didn’t set out to build a company. But I ended up dropping out of college and going all-in. That’s where the story really begins.

Stacey Epstein (02:43)
You wouldn't be the first tech guy to start a political party—maybe that's still in your future! But let’s go back: what was that first idea?

Quinn Favret (02:54)
At the time, I was working on a project for a campus restaurant. They had no idea how many customers or how much food they’d need week to week. They were either overstaffed with no customers, or overwhelmed with too many. So we built a demand forecasting model that could predict, with 95% accuracy, how many burgers they'd sell at 3 p.m. on a Friday—based on things like weather, traffic, and the campus calendar. That was the first real idea.

Stacey Epstein (03:33)
Interesting. And is Tavus the next big idea?

Quinn Favret (03:46)
Exactly. That first one was smaller. Tavus is the big idea. We sold the first company, and I started thinking about what was next. My co-founder and I—who’s also a close friend—started talking about how people communicate with technology. How do we talk to machines? I’ve always been frustrated by how restrictive computers can feel—like they’re in the way, rather than meeting you where you are.

So we asked: what if machines could see, listen, speak—what if they could act more like humans? That’s what led to Tavus.

Stacey Epstein (04:21)
Mm-hmm.

Quinn Favret (04:43)
We’ve spent the last four or five years building what we call an operating system for human-AI interaction. It’s real-time, back-and-forth conversation with a digital human—or “replica”—that sounds and feels like talking to a real person.

Stacey Epstein (05:07)
Interesting. I can think of a million use cases. What are the main ones you see today?

Quinn Favret (05:14)
It really comes down to: where would you love to have a human, but can’t—because of cost, scalability, or availability?

A big one is AI therapy. There’s a massive shortage of therapists, and most people can’t afford or access one. So we see use cases in:

1. Healthcare – patient intake, AI nutrition coaches, companionship.
2. Education – from corporate L&D to K–12 tutors.
3. Recruiting – running more efficient and thorough interview processes.

These are all areas where traditional AI hasn’t delivered what humans can. But a digital human can.

Stacey Epstein (06:29)
Yeah, I’ve talked about this on the show before—my daughters are taking French, and I can’t help them at all. I told my daughter, “Just ask ChatGPT!” She was surprised at how helpful it was. The beauty is you can ask the dumbest question 10 times and it doesn’t judge. It’s such a great tutor for that reason.

Quinn Favret (07:23)
Your daughter has already made it further than me—I didn’t make it past French I. But totally agree. We shouldn’t have to adapt to machines—they should adapt to us. That’s what makes digital humans so effective. They understand the nuances of communication—tone, gestures, eye contact—things that make learning more natural.

Stacey Epstein (08:17)
My audience is mostly marketers and go-to-market executives, so let’s talk about that use case. How can this be used in business to make us better marketers and sellers?

Quinn Favret (08:43)
We’re seeing great traction in digital recruiting—AI-powered screeners and interviewers that improve candidate experience and surface better signals.

For go-to-market, we’re seeing a ton of innovation. Imagine an AI sales rep that pops up on your site, answers questions, books meetings, and follows up. Or a digital CSM who truly knows each user in a PLG motion. We’re enabling deeper engagement across the funnel—from initial interest all the way through customer success.

Stacey Epstein (10:21)
We’ve had OneMind and Qualified on the show—both in this space. It sounds like you’re not an AI SDR tool, but more of a foundational platform that others can build on. Is that fair?

Quinn Favret (10:51)
Exactly. We’re not building application layers like an AI SDR. We’re building the infrastructure—the digital human models, visual representation, perception, text-to-speech. We white-label our APIs so others can build on top. If OneMind wanted to add digital humans, they could use our system instead of building it all themselves.

Stacey Epstein (11:45)
That makes sense. But for marketers, that can feel intimidating. AI is already overwhelming, and it's easier to just buy a tool like Qualified than to build from scratch. What does it take to use Tavus? Are you mostly selling to big companies?

Quinn Favret (12:34)
We sell to a wide range—from 10-person startups to Amazon and Alibaba. Sometimes we help them build a specific use case, like a digital coach or AI recruiter, which they then offer to end users.

But we know we need to make Tavus easier to use. Soon, you’ll be able to say, “I want a digital human who looks like this, does that, and serves this role,” without needing engineering support. That’s the direction we’re going.

Stacey Epstein (13:52)
It’s early days, but this reminds me a bit of the Salesforce ecosystem. You could buy products, or build your own apps on top. The possibilities now with AI are even broader—and it sounds like Tavus makes it easier to spin things up.

Quinn Favret (14:41)
Totally. Internally, when we need to hire or launch something new, we always ask: why can’t we use a Tavus human to do this? That mindset has opened up a lot of exciting use cases for us.

Stacey Epstein (15:18)
I was texting with friends about this—one said companies will eventually run with just one person. And I said, yeah, but original thought still comes from humans. This podcast wouldn’t exist if I hadn’t had the idea to educate marketers on AI. AI helped me launch it—Riverside, Buzzsprout, editing—but the vision was mine.

Quinn Favret (16:02)
Exactly. The way I work today is so different than even a year ago. AI gives me tools to be more effective, more creative. But yes—the noise is real. There’s so much happening, so fast, it’s hard to know what matters. But that pace of innovation is also what makes this moment so exciting.

Stacey Epstein (17:02)
Totally. Some people say AI will take all our jobs—but that misses the point. The original ideas come from us. AI just helps us get them out into the world faster and better.

I’m a writer. I use AI to write all the time. And sure, sometimes I think I should resist it—but if it helps me express my ideas more effectively, why wouldn’t I use it?

Quinn Favret (18:39)
Yes! AI is an extension of us—not a replacement. Work will change, just like it did in the Industrial Revolution. But that’s not a bad thing. Humans weren’t built for monotonous fluorescent-lit tasks. We’re creative. We’re original. That’s our superpower.

Stacey Epstein (19:57)
I 100% agree. We started with: "What can a human do that AI can replicate?" But I love that we’re ending with the importance of humans in directing all of this—using our strengths to make the world better, faster. Full circle, back to your point about politics.

Quinn Favret (20:25)
Exactly. That’s what excites me about Tavus. We’re not just building point solutions—we’re building a platform that gives people the tools to have a ripple effect. That’s how we make real, positive change.

Stacey Epstein (20:45)
I love that. A full-circle moment. I’m glad you’ve found your path. So how do we find you?

Quinn Favret (21:02)
Check out tavus.io. There’s a live demo you can try—it’s magical. I probably didn’t explain it perfectly here, but once you experience it, you’ll get it. We have free self-serve plans, or you can reach out to me directly and I’ll help you get started.

Stacey Epstein (21:29)
Awesome. I can’t wait to try it out. Thanks so much for joining the show.

Quinn Favret (21:34)
Amazing. Thanks, Stacey.