Investment research firm Morningstar evaluates companies using a concept called an economic moat. It’s borrowed from medieval defense strategy: the wider and deeper the moat around your castle, the harder it is for anyone to breach the walls and burn your keep. 

In investing terms, a company’s moat is whatever structural advantage keeps competitors from stealing its business. It might be structural: a proprietary supply chain, regulatory barriers, deep institutional expertise, or switching costs so high that customers never leave. Or, your moat may be rooted in the service your company provides and the expertise, accountability, and trust that generates. As executive product coach Rich Mironov ponders in his excellent piece on this subject, “what moats can we construct other than time-to-market to defend our products and deliver ongoing value?”

The wider the moat, the more durable the business. No moat, and you’re one well-funded competitor away from irrelevance. In the upcoming AI era, we argue that your service moat may prove to be your company’s biggest strength.

The SaaS Reckoning.

For the past two decades, software-as-a-service companies have built their entire economic model on one beautiful premise: software has near-zero marginal cost. Build it once, sell it to a million people, and every new subscriber is almost pure profit. That’s why SaaS valuations have been astronomical. The product scales infinitely while the cost stays flat. 

AI is about to invert that logic. 

The cost curve of AI-generated software is heading toward zero. Compute is expensive today, but so was memory when a 16K-to-64K upgrade cost a thousand dollars and a 5MB hard drive would see you two thousand out of pocket. Today you get a million times the storage for a quarter of the price — and yet, more hard drives are sold now than ever before. 

The same pattern will play out with AI-assisted development. The cost per unit of output will plummet, while total volume of output explodes. There will be more software built, more products shipped, more systems deployed than at any point in history. The size of the custom software market is expected to nearly triple by 2030. And every single one of these new products will need someone to validate that it works, that it’s secure, that it does what it claims to do. 

Consider what happens when a reasonably capable person can sit down with an AI model and say, “Build me an expense tracker that does X, Y, and Z.” What is delivered is not a toy, but a functional application, tailored to their specific needs, and built in an afternoon. They don’t need your subscription. They don’t need your UI. They don’t need your brand.

Think about every SaaS tool where the product is the service, like expense trackers, project boards, and note-taking apps. The entire value proposition is surface-level, and there’s no expertise behind the curtain, no guarantee beneath the workflow, no human accountability if something goes wrong. That’s not a moat; that’s a feature set. And feature sets are now a weekend project.

The companies most at risk are the ones that deliberately drained their service moats in pursuit of higher margin. They optimized for scalability and got fragility instead. 

How to Widen Your Moat Right Now.

The service you provide, what we call the service moat, is your biggest opportunity in this landscape. It’s where most companies either survive or fool themselves. 

What clients of service-moat companies are buying is the promise that things won’t go wrong — and if they do, a human being is accountable. If you’ve ever seen the penalty structure for late tax payments, you understand why that promise is worth real money. Clients are buying human expertise encoded into a system of guarantees that no AI can independently provide. Guarantees require accountability, and accountability requires a human on the other end willing to stake their reputation on the outcome. 

This principle extends across every industry, from fintech to healthcare. The moat is migrating away from the technology and toward the judgment layer that sits on top of it. The question is no longer “can we build it?” The question is “who stands behind it?” In this world, your mobile app and web interface are not your moat. Those are delivery vehicles, and increasingly, they’re commoditized ones.

Apply the Moat Test to Your Business.

Every organization should be stress-testing its roles, its products, and its business model against a simple filter: 

If AI can replicate it, it’s not a moat. 

That means looking at every function in your organization and asking four questions about each task within it: 

• Can AI do this entirely? 
  If yes, stop assigning humans to it. Automate it and reallocate the talent. 

• Can AI do part of it? 
  If yes, redesign the role around the human judgment that remains essential. The person owns the outcome; AI contributes to execution. 

• Can AI make this faster? 
  If yes, raise your throughput expectations accordingly. Same cognitive work, higher velocity. 

• Can a human add meaningful quality to AI’s work? 
  If yes, allocate your best and most creative contributors to this problem. 

• Does this require trust, ethics, creativity, or accountability? 
  If yes, this is your keep. Protect it. Invest in it. Build your value proposition around it. 

The companies that survive the next five years will be the ones that can clearly articulate what their humans do that AI cannot; and build their pricing, their brand, and their organizational structure around that answer. 

Weathering the Siege.

In a world where software is rapidly becoming cheaper to produce, widening your critically important service moat is paramount. Thinking carefully about your service moat cuts through the hype and the panic alike and forces a single, actionable question: how do you build the trust your customers have in you, how do you provide guarantees of quality, and how can your teams members’ expertise be applied creatively to new challenges, using new tools? 

If you can answer these questions clearly, you know where your keep is — and AI becomes a faster drawbridge to reach it. If you can’t, you’re left standing behind a crumbling wall… and the siege engines are already rolling in.