An AI startup for plumbers just hit $1B

In partnership with

Good Saturday.

A Fortune piece dropped on Monday that I haven't been able to stop thinking about.

Avoca raised $125M+ at a $1B valuation. The product? An AI voice agent that picks up the phone for plumbers, HVAC operators, and roofers when their customers call.

That's the entire pitch.

Here's the line from Kleiner Perkins partner Leigh Marie Braswell that stopped me: "This is definitely an industry that's been overlooked by Silicon Valley. I think people don't realize how big it is or, quite frankly, how important it is."

She's an early Scale AI employee. She led Avoca's Series A. The HVAC industry alone is a $50B market.

While every founder I know spent the last 18 months building yet another AI writing tool, two cofounders quietly built a unicorn answering phones for plumbers.

Three lessons in here for anyone in the NTE audience trying to figure out what to actually build.

Lesson 1 — "Unsexy" is the entry ticket, not the disqualifier

Avoca sells to operators most SF founders have never met. HVAC dispatch managers. Roofing company owners. The people whose business cards you'd never collect at TechCrunch Disrupt.

That's not a bug. It's the whole reason the round got done at $1B.

When a category gets called "unsexy" by SF, what's actually happening is a coordination failure. Smart founders avoid it. Smart investors follow them out. The buyers don't care what label SF puts on their industry. They care whether their phone gets answered at 7pm on a Saturday when the basement is flooding.

If you find yourself dismissing an idea because it sounds boring, that's a signal to dig in, not drop it. The label is what's keeping the category un-built.

Customers in "boring" industries also pay faster, churn less, and require less product polish than the SaaS audience the average founder is fighting over.

The reader of this newsletter who works in or near one of these industries already has an unfair advantage. Most don't realize it.

Your clients can't tell you dictated that.

That polished proposal? Dictated in two minutes. The detailed follow-up email? Spoken between meetings. The Slack update with clean formatting and exact numbers? Voice.

Wispr Flow turns everything you say into professional, send-ready text. Speak naturally — go on tangents, change your mind, circle back — and get output that reads like you spent twenty minutes writing it.

89% of messages sent with zero edits. Millions of professionals use Flow daily. Works in every app on Mac, Windows, and iPhone.

Start flowing free

Lesson 2 — Vertical workflow beats horizontal capability

ElevenLabs ships voice. Vapi and Retell ship the conversation engine. Bland ships the orchestration. The components for "an AI that talks on the phone" have been commoditized for over a year.

Most founders looked at that and thought: cool, the building blocks are cheap, what's left to build?

Here's what's left.

The difference between a generic voice agent and Avoca is that Avoca knows the difference between an emergency drain backup and a routine water-heater quote. It reads ServiceTitan inventory before promising same-day service. It escalates correctly when a tech is double-booked.

That's not a model capability. That's a workflow capability.

If your AI startup pitch is "we're like ChatGPT but for X," you don't have a business yet. If your pitch is "we know the seventeen specific decisions a dispatcher makes every shift, we automate the boring fourteen and escalate the spicy three," you might.

Lesson 3 — Plug into the system of record, don't try to replace it

Avoca didn't try to replace ServiceTitan. ServiceTitan is the boring CRM that runs the trades industry. It's what every dispatcher has open all day. Nobody's switching off of it.

So Avoca built the voice layer that talks to it.

This pattern is worth memorizing. The boring SaaS that already won a vertical a decade ago is the new operating system. Anything built on top of it inherits the gravity, the data, and the customer relationship.

ServiceTitan for trades. Housecall Pro for smaller shops. Procore for construction. Toast for restaurants. Mindbody for fitness. Each one is a wedge for someone who shows up with a workflow agent that integrates well.

Stop trying to build a new system of record. Start building the agent that lives on top of one.

Where this connects to what we do at NTE

This is the exact pattern NTE Pro is built around. Most of the 7,000+ ideas in there aren't "build a new ChatGPT competitor." They're real workflows in real industries that nobody's automated yet, with the buyer already sketched out.

You're not supposed to pick one off the shelf. You're supposed to read fifty of them until your pattern recognition gets sharp enough that you walk into a friend's HVAC shop and within ten minutes have spotted three things they're paying humans to do that they shouldn't be.

That's the muscle Avoca's founders had. You can build it.

For the B2B builders: Avoca's customers include 1-800-GOT-JUNK?, Goettl, and Turnpoint Services. The next Avoca is sitting in someone's hiring page and someone else's state filing right now. WhoFiled is where you see those signals months before the Fortune piece runs.

One question for your weekend

What's an industry you have direct access to, through family, friends, or your own work history, that Silicon Valley would call "boring"?

That's your unfair advantage. Most of you have one and don't realize it.

Spend 30 minutes today listing the top five workflows in that industry that still run on phone calls, spreadsheets, or paper. Pick the one with the most repetition.

That's your Avoca.