Conversation
Highlights

How Augment Code Achieved 80%+ Win Rates in AI’s Most Hyped Category

Matt McClernan joined Augment Code as CRO in November 2024, walking into one of the most hyped markets in tech history. Every week brought new funding announcements, product launches, and competitors claiming breakthrough AI coding capabilities.

Within nine months, Augment’s competitive win rate jumped from 50-60% to north of 80%. In a recent episode of Unicorn Builders, Matt—who transitioned to CEO in July—shared the specific GTM decisions that drove this acceleration.

The 90-Day Problem: Finding Signal in Noise

The AI coding space presented a unique challenge. Unlike mature categories with clear buying criteria, enterprises were still figuring out what “good” looked like. “You step into the arena in a place that is filled with hype,” Matt explained. “There are many of the products we’re competing against are focused on individual developers. And you’re hearing every day of new funding rounds of new companies that are entering the space.”

Augment had invested years building infrastructure for codebase understanding—fundamentally different technology than the code generation tools dominating headlines. But technical differentiation alone doesn’t create a repeatable sales motion.

Matt’s first 90 days focused on three parallel workstreams: “Figuring out where we fit in the space, who our customers are, what’s the ideal sales process and experience like for our customers to buy from us and then quickly forming this flow of information to our engineering and research team so we could double down in the areas that were winning.”

ICP Design as Differentiation Validation

Most early-stage companies cast wide nets, optimizing for volume. Augment did the opposite. “We had to orient for customers who had a code base that would help our products shine, where we would really be differentiated,” Matt said.

The initial ICP criteria:

North American companies (operational focus for small Palo Alto team)

Organizations already using competing AI coding products

Engineering teams working with complex, multi-year codebases

That second criterion proved critical. They specifically targeted companies with existing AI coding tool experience—buyers who could make informed comparisons. “So they had a palette, they had some sort of experience they could compare us against,” Matt explained.

This wasn’t about finding easy wins. It was about finding buyers sophisticated enough to understand why codebase context matters at scale. Companies using simpler tools would eventually hit limitations. Augment needed to be there when that happened.

Engineering Feedback Loops as Competitive Advantage

In hyper-competitive markets, velocity of learning trumps perfection of execution. Matt built direct pipelines from customer conversations to the engineering team. “We had to understand if there were patterns that we’re hearing from customers of gaps in our product, what those would be to help us go and win more consistently.”

This created a compounding effect. Each sales conversation generated product insights. Product improvements shortened sales cycles. Shorter cycles meant more conversations and faster iteration.

The validation came quickly: “Fortunately we found very early on consistent success and so much so that as we’ve expanded our ICP, our win rate has actually accelerated or increased.”

Going upmarket didn’t dilute their advantage—it amplified it. “The farther north we go in terms of the size of these software engineering teams, the more apparent that is,” Matt said, referring to their codebase understanding differentiation.

The Market Education Playbook

Matt’s time at MongoDB shaped his competitive philosophy. When hyperscalers launched competing database services, the initial reaction was existential threat. The reality proved different. “We found early on that they actually were validating the market opportunity,” he explained. “And over time, as they saw us continue to win at a higher and higher rate in competitive scenarios and otherwise, the market sort of started to come to us.”

He’s seeing identical dynamics with frontier labs and AI coding startups. “I think it’s actually all to our benefit,” Matt said about the competitive landscape.

The pattern: “The majority of the companies we compete against, they built products that are optimized for producing code with AI facilitating, we call that zero to one projects.” These tools excel at generating simple applications from natural language—impressive demos, but limited production utility.

“Those types of products, they struggle when they’re used day to day in large code bases,” Matt explained. “What we’re seeing is companies on this maturity spectrum as they bring in these types of products and they realize the developers stop using them because they get in the way. They also start to realize they’re polluting their code bases.”

This creates Augment’s entry wedge. Competitors aren’t stealing share—they’re qualifying leads. “It’s like they’re forming a pallet and they’re able to become more discerning. And that’s where Augment steps in.”

Hiring for Ambiguity Over Execution

Scaling in zero-precedent markets requires different seller profiles. “Some companies might look for salespeople who are focused on making a lot of money and closing big deals,” Matt said. “What’s really important here is finding sellers that of course are very successful in their careers prior to joining, but are very mission oriented.”

The specific attributes: “They need to be motivated by how big this market and opportunity is. Being able to represent a transformational piece of technology and also understand coming in and that there are no guarantees, there’s no precedent for what we’re doing here.”

Standard enterprise sales playbooks don’t apply when buyers are still figuring out evaluation criteria. Augment needed sellers comfortable navigating uncertainty, synthesizing technical feedback, and collaborating with engineering teams to shape product direction.

“We’ve oriented around these types of sellers that understand the risk that’s at stake here, but they’re motivated by the larger opportunity to go and figure something out,” Matt explained.

The Enterprise Maturity Paradox

One insight surprised Matt: enterprise AI adoption shows zero correlation with company sophistication. “There’s one company that’s north of 100 billion market cap that’s based out here and they are a leader in their space in technology. Incredible team of engineers, thousands of engineers. We were the first AI for code product that they’ve adopted and this was just a few months ago.”

Meanwhile, similar companies were “three years into their journey and they’re on the third or fourth iteration” of AI coding tools.

This variance eliminates assumptions. “You don’t know until you meet with a customer where they are in the journey,” Matt said. Some Fortune 100 companies are just starting. Some mid-market tech companies are in replacement cycles.

The GTM implication: discovery can’t be templated. Each conversation requires diagnosing where that specific organization sits in their AI adoption arc. “You need to start with where the customer is and then help them understand where your product, what it’s going to deliver for them to take, meet their goals and outcomes that they’re seeking.”

CRO Hiring: Operator First, Executive Later

Matt’s transition from CRO to CEO in July shaped his perspective on revenue leadership. His advice to founders: “I’d say earlier, is actually better.”

But not just any CRO. “You want someone that’s willing to step in and they should strive to be the best IC the best seller in the company. They can’t scale an organization successfully unless they’re doing it.”

The specific behaviors: “They’re having these conversations with customers daily. They are failing. They’re sharing their experiences with their team and allowing them the opportunity to coach them so they can create that sort of an atmosphere.”

This inverts the typical executive hire. Instead of bringing in a scaled-company operator to implement proven playbooks, early companies need revenue leaders who will discover the playbook alongside the team. “Find someone who’s willing to be humble, who wants to learn, who’s mission driven, who’s and is willing to gradually earn the opportunity, then scale out an organization once they’ve helped you find the product market fit.”

The title comes after the work, not before.

The Durability Question

Augment’s win rate acceleration validates their thesis: enterprises need AI coding tools purpose-built for production codebases, not demo applications. As Matt observed, simpler tools create baseline expectations that sophisticated buyers eventually outgrow.

The strategic question isn’t whether to compete in hyped categories—it’s whether your differentiation compounds or commoditizes as the market matures. Augment bet that codebase understanding would matter more over time, not less.

The 80%+ win rate suggests they were right.