Article 3: Replication is not displacement

If I had to pick one insight from the last two years of watching AI reshape the software industry, it would be this: people consistently overestimate how much easier AI has made it to displace an incumbent software business, while correctly observing that it's made it easier to replicate one.

These are not the same thing. And the gap between them is where a lot of the interesting investing happens right now.

What's actually gotten easier

Let's be honest about what AI has genuinely changed. Building a product that looks and functions like an existing vertical software product is materially easier than it was three years ago. The UI can be generated. The core workflows can be modelled. A prototype can be convincing in a demo. For a well-resourced challenger, getting to 'plausible competitor' stage is faster and cheaper than it's ever been.

This is real, and it would be naive to wave it away. In categories where the product is the primary differentiator, where customers choose based on features, design, and price, this compression of build time does meaningfully lower barriers to entry.

But most vertical software markets are not like this. And this is where the replication versus displacement distinction matters.

Why displacement is still hard

Consider what it actually takes to displace an incumbent in a vertical market.

First, you have to find the customers. Vertical markets are often tight-knit communities: industry conferences, trade associations, word of mouth networks that have been running for decades. The incumbents are embedded in these networks in ways that don't show up in any product comparison. They sponsor the conferences. Their founders sit on industry boards. Their salespeople have relationships with buyers that are years or decades old. Getting in front of the right people, at the right moment, with enough credibility to be taken seriously: that's a go-to-market problem rather than a product problem.

Second, you have to earn the trust to be considered. Switching software in a vertical business is not a trivial decision. It disrupts operations, requires staff retraining, and carries real risk of something going wrong during transition. Customers in regulated industries are especially risk-averse, and they'll sit on an older, less feature-rich product for years rather than go through a migration. The bar for 'good enough to switch to' is much higher than 'better than the alternative.'

Third, you have to execute the switch. Even if a customer decides to move, the implementation process is where a huge number of challenger companies fail. Migrating data, training users, handling the edge cases that only emerge in real-world use, managing the anxiety of an operations team going through a system change: this requires experienced customer success people who understand the industry. You can't replicate that with a better codebase.

Go-to-market is the real moat

Here's what I keep coming back to: in a world where product differentiation is compressing, the businesses with the most durable advantages are the ones with the best go-to-market engines. The relationships, the referral networks, the reputation, the customer success capability: these are the things that actually determine whether a business grows and retains customers, and they're all things that take years to build and can't be shortcutted by better technology.

This is a partial inversion of how a lot of tech investors have historically thought about moats. The assumption has often been that product is the primary driver of competitive advantage, with go-to-market as the execution layer that monetises it. In vertical SaaS in an AI-native world, I'd flip that: go-to-market and customer success are the primary moats, and product quality is the table stakes that gets you into the conversation.

Customer success deserves particular attention here. The best vertical SaaS businesses I've looked at have customer success teams that function more like industry consultants than software support agents. They understand the customer's business deeply. They proactively surface value. They're the reason a customer who had a bad quarter with the software doesn't churn, because there's a person they trust who helped them through it. That kind of relationship is extraordinarily hard to displace, regardless of how good a competitor's product is.

The punchline

The businesses that are going to look most vulnerable in the AI era are the ones that competed primarily on product features and assumed that was enough. The businesses that are going to look most durable are the ones that built deep market presence, strong customer relationships, and excellent customer success, and treated product quality as a baseline expectation rather than a differentiator.

In other words: the fundamentals that great software investors have always cared about, retention, relationships, market depth, team quality, are more important now than they've ever been. The hype cycle wants to tell you that everything has changed. The truth is closer to: the things that always mattered, matter more.

Author: Yousif Hanna