One founder sat in the living room, taking a call. The other crawled into the bathroom, closed the door, and sat on the floor to take a second call. Why? So the clients on the other end would believe they were in two separate conference rooms — not two guys sharing a single apartment.
This isn’t a startup meme. It’s the real story of Decagon’s early days, told by co-founder Ashwin Sreenivas on stage at a16z’s San Francisco office. And it’s the most honest, unglamorous, and brutally useful lesson on product-market fit I’ve ever heard.
Decagon now builds enterprise AI customer service agents. Their clients include airlines, telecoms, and banks. The team has grown from 30 people last year to over 500 today, on track to hit 1,000 by year-end. But four years ago, it was just Ashwin and his co-founder Jesse, hiding in a bathroom to sound bigger than they were.
Here’s what they did right — and what almost every other startup gets wrong.
Step 1: Don’t Start With a Differentiation Strategy
Most founders obsess over how their product is different. Ashwin and Jesse did the opposite. They started with a buyer persona: someone who has real budget and real knowledge of the market.
“Knowledgeable buyers are dangerous,” Ashwin said. “If you sell to someone who has never been pitched before, you get a false sense of security. They buy because you’re the first, not because you’re good. Eventually, you’ll hit a buyer who knows the landscape — and they’ll tell you the seven other options that are better than you.”
So they deliberately went after buyers who had already been pitched by dozens of competitors — but hadn’t bought yet. That’s the sweet spot. Those buyers can tell you exactly why every existing solution fails. And if you solve those specific failures, you don’t need to fabricate a differentiation story. You just repeat what the customers told you.
The One Question That Changes Everything
Ashwin shared a simple, brutal question they used for early customer discovery: “What are the top two or three things you want to get done in the next six to nine months?”
Don’t ask if your idea is important. Everyone says yes to that. Ask for their real priority list. People only spend money and time on the first two items. Everything else is noise. If your solution isn’t in their top two, you’re selling to a ghost.
They also used a clever validation trick: after collecting a list of reasons why previous solutions failed, they’d go back to the same buyers and say, “If we fix A and B, will you pay us $80,000?” They only started building after three customers said yes.
The Paradox of Resource Scarcity
When you’re two people in a single apartment, you can’t afford to build anything that doesn’t directly lead to a signed contract. That scarcity, Ashwin argues, is a feature, not a bug.
“Customer obsession wasn’t a value we picked out of a book. It was forced on us by having no resources. Sales can only sell what customers actually want. Engineering can easily drift into building cool things nobody asked for. But when you’re stretched thin, you only build what customers explicitly say ‘I won’t sign without this.’”
This is where the bathroom story comes back. Faking a bigger office wasn’t a gimmick — it was a survival tactic. They had to look credible to close deals. But the real lesson is that those constraints forced them to listen harder than any well-funded competitor ever would.
Why You Should Never Convince a Customer
Ashwin said something that should be framed on every founder’s wall: “If you have to convince a customer, they don’t have the problem.”
When a problem is real and painful enough, the customer is the one pulling you forward. They’re desperate for a solution. Your job isn’t to persuade — it’s to prove you can deliver. If you find yourself constantly justifying and arguing, you’re probably solving a problem that doesn’t exist.
This reframes the entire sales process. Instead of selling, you’re screening. Is the pain real? Is the budget real? If yes, the rest is just execution.
The Twist: Competitors Are a Good Sign
When asked about the wave of competitors flooding the AI customer service space, Ashwin smiled. “A truly good market — one where buyers have real pain and real money — will always attract competitors. There’s almost no good market that has zero competition.”
His advice: don’t worry about competitors unless you’re losing deals to them. If you’re not losing, ignore them. If you start losing, find the specific reason and fix it. That’s it. No obsession, no panic.
Speed Isn’t a Management Trick — It’s Debt to Customers
Everyone asks how Decagon moves so fast. Ashwin’s answer: “We never give ourselves a deadline cushion. If a customer says two days, we deliver in two days. We don’t extend it just because we’re busy. That pressure is what creates speed.”
It’s not a process. It’s a commitment. And it’s the only thing that scales when you’re going from 30 to 1,000 people.
The Real Secret to Hiring
Early on, they hired for two things: tolerance for chaos and extreme technical competence. Why? Because in a five-person company, the sixth person will evaluate everyone else. If they don’t respect the existing team, they won’t stay. And if they can’t handle shifting priorities (a project you worked on for four days gets killed), they’ll burn out.
Ashwin’s only regret: not hiring faster. They were too conservative with cash, which slowed growth. But the lesson is clear — speed of hiring is a competitive advantage, as long as you maintain the bar.
Culture Is Not Organic — It’s Transplanted
With offices in San Francisco, New York, and London, Decagon forces every new hire to spend two weeks in SF first. They send veteran employees to new offices to seed the culture. Ashwin explains why: “Culture is how people work together. If one team expects instant response and another team queues everything, they’ll clash. You can’t let culture grow wild — you have to intentionally transplant it.”
The Moment That Made It Real
When asked what week he’d relive, Ashwin didn’t pick a funding round or a product launch. He picked the week they won Eventbrite as a customer. It was a bake-off: both products were shown to users alternately. They’d refresh the screen, see a user query, and either answer it themselves or watch the competitor’s answer. They’d adjust in real time. That raw, high-stakes, founder-in-the-trenches moment — that’s what he’d relive.
Because that’s where the real product-market fit is found. Not in a boardroom with a slide deck. In a living room, with one founder on the floor of a bathroom, and a competitor’s answer popping up on the screen.
Stop pretending your roadmap is precise. Stop fabricating differentiation. Start asking the question that matters: Is the problem real, and does the customer have the budget to solve it?
Everything else is just noise.
FAQ
Q: Isn't faking a larger office dishonest?
A: It's a survival tactic when you're resource-constrained. The key is that they were actually solving real problems for clients — the fake office was just a psychological crutch to get the meeting. The product delivered.
Q: How do I know if a customer's problem is real enough to build for?
A: Ask them what their top 2–3 priorities are for the next six months. If your solution isn't one of them, move on. Then ask: 'If we fix this, will you pay us X?' Get a verbal commitment from at least three customers before writing a single line of code.
Q: What if my market has many competitors — should I pivot?
A: No. A market with many competitors is often a sign that the problem is real and the money is there. Only worry if you're losing deals to them. If you're not, keep your head down. If you start losing, find the specific reason and fix it.