In 2018, my wife and I were shopping for a mortgage. I put my phone number into Lending Tree and within seconds — not minutes, seconds — I had 18 missed calls. No context. No quote. No "let's understand what you're looking for." Just call after call after call hitting my voicemail.
Around the same time, I was buying a car entirely over text. A guy in the next city over. We negotiated the price, figured out pickup, handled the whole thing through SMS. And I remember thinking: "this feels right. That other thing felt extraordinarily wrong."
If I can buy a car over text, I should be able to buy financial products over text.
That question is why Mav started. But the real decision, the one that's defined everything since, was choosing to build exclusively for insurance. There's a question I come back to whenever things get complicated: who do I serve? When you chop the layers apart, it makes any decision easy. We serve insurance agency owners. Their best interest drives every decision we make. And going all in on their world was the first expression of that.
We have competitors who stay horizontal. One AI layer stretched across mortgage, life and health, home services. The idea being that call center technology should apply to everything. I get the logic. I just think it caps how much value you can actually deliver.
There is extraordinary nuance to what it means to operate in the insurance space. The difference between captive and independent agencies. Standard versus non-standard. The way an Allstate or Farmers agent in Texas buying leads from Everquote needs to follow a different ruleset than an independent agent working aged leads in Ohio. Knowing those nuances changes everything we build: the conversational flows, the qualification logic and the routing logic, to name a few.
And years of data in one lane compounds. We may not have the most groundbreaking, fine-tuned model, but we have some of the deepest vertical technology in insurance right now. That depth is something no horizontal player can shortcut, and I think it's going to pay dividends in the next chapter.
We track an internal benchmark: the transfer rate a class-A human call center achieves on the same lead flow and agency type. That's the gold standard. We're now matching, and in some cases, surpassing that number.
Cost is what gets us in the door. We operate at a fraction of a human call center, and the margin math only gets more compelling at scale. But performance is what keeps our customers. You don't become higher performing unless you're delivering a better experience. More utility. More of the funnel captured. Less leakage.
Here's the thing most people aren't thinking about yet: the moment AI consistently outperforms a human call center, the horizontal AI call center becomes a commodity. If everyone can spin up a generic AI dialer, the differentiator isn't the AI itself - it's what the AI knows.
That's the future we're most excited for.
The foundation models are getting so good, so fast. There's going to be a time, sooner than most people realize, where your personal AI just knows who you are, what coverage you have, when your policy expires, and it surfaces the right options before you ever open a browser. That world is closer than you think.
When that world arrives, the agencies plugged into that ecosystem win. The ones that aren't get left behind.
We're shipping some of the most ambitious product we've ever built in the back half of this year - things that only a vertical AI built for insurance can pull off. Not just top-of-funnel performance improvements, but capabilities that reshape how an agency actually operates. We can build this because we have the data, the integrations, the domain depth, and years of compounding context that come from serving one industry and one industry only.
Our job at Mav is to democratize AI for insurance agencies. And to make sure that when the next wave hits, our customers are right there in the mix.
That's who we serve. That's why we went all in on insurance. And we're just getting started.
