Your Moat Is Leaking and Why Micro-SaaS Will Win the AI Era
December 2024. Satya Nadella goes on the BG2 podcast and declares that SaaS is dead.
I watched that podcast. I tweeted my thoughts on it at the time:
"Very interesting take and certainly thought-provoking. I think the future would perhaps look something like this: AI agents will build the backend based on business needs (in real-time) and then actively monitor business changes to continue refining the business logic (in the backend). This is more plausible than the AI agents are 'the backend'. Business logic would need to be predictable 100% of the time; AI, as we have it today, is not!"
— @sarmaad, December 20, 2024
At the time, AI agents were just getting started. AI coding was in its infancy. AI was, for most people, an elaborate chatbot. ChatGPT was all the hype.
Fast forward 14 months, and the $1 trillion wipeout in software stocks has made it the dominant conversation in every SaaS boardroom. Ready or not.
AI Is Dissolving Your Moat
Historically, the value proposition of SaaS came down to two economic realities.
First, building software was prohibitively expensive and slow. Most small and mid-sized businesses simply could not justify the cost or the timeline. The ROI did not stack up.
Second, software projects fail at an alarming rate. The risk of investing six figures into a custom build, only to see it collapse mid-delivery, was a dealbreaker for most businesses.
SaaS removed both barriers in one move: predictable subscription pricing, maintained by someone else, with the risk sitting on the vendor's balance sheet, not yours.
That was the moat. And for two decades, it held.
Now AI is dismantling both pillars simultaneously. When the cost of building software drops 100x, and it has, the entire economic case for SaaS does not just weaken, it collapses. Businesses that could not afford to build can now build. Projects that used to require 18 months and a team of engineers now take weeks. The vendor's risk absorption was valuable when building was hard. When building becomes cheap, fast, and increasingly reliable, what exactly are you selling?
But My SaaS Is Different!
That is probably what you are thinking right now. "This does not apply to me, right?"
To be fair, some SaaS businesses are genuinely defensible, those built around regulatory compliance, government licensing, auditing requirements, or special permits. The switching cost there is not convenience; it is legal exposure. AI does not dissolve that overnight.
But that is a small club. Most SaaS businesses do not operate behind regulatory walls. And if your defence is anything other than "our customers literally cannot leave without risking a fine or losing a licence", you need to keep reading.
Let me address the three most common arguments:
"We have years of accumulated customer knowledge"
AI models do not just match your accumulated knowledge; they synthesise it across every industry, every use case, and every customer type simultaneously. What took you a decade to learn, an AI-native competitor can inherit on day one. Your knowledge just became a commodity rather than a moat.
"Our product is too complex to replicate"
Complexity was protection when the cost of building was high and timelines were long. Now, complexity is a red flag. Customers are paying for outcomes, not complexity. An AI-native competitor can deliver the same outcomes with less friction. The complexity of your SaaS just becomes a reason customers churn.
"We integrate with hundreds of tools and APIs"
This is the most dangerous one to hide behind. AI agents are API-native by design. Connecting systems, orchestrating workflows, and moving data between tools is precisely what they are built to do. The integration layer you spent years building is exactly what AI agents do out of the box.
Here is the thing. I am not writing a eulogy for SaaS businesses. Far from it.
I want to send a wake-up call to founders who are paying attention. The same AI that is eroding your moat can also strengthen your position. The founders who move fast, embed AI deep within their product, and rethink what value they actually deliver to their customers, will not only survive this shift, they will dominate it.
Big SaaS Got Fat. Time to Trim It Down.
The problem with big SaaS is not that it is big. The problem is that size became a substitute for intimacy.
When building software was expensive, customers accepted being one of thousands. They tolerated feature requests disappearing into a backlog. They accepted a roadmap that never quite fit their business. Because the alternative, building their own, was worse.
That tolerance is running out.
Traditionally, the bigger the SaaS, the harder it was for customers to leave. Size signalled stability, breadth of use cases, and trust. The conversation often went something like this: "They have hundreds of customers, they must be doing something right. They can accommodate whatever my business needs."
And that logic held, because building software was costly and slow. A large SaaS absorbing that complexity on your behalf was genuinely valuable. The bigger the product, the more problems it solved, the stickier it became.
But that calculus is changing, and changing fast. When the cost of building collapses, size stops being a signal of quality and starts being a signal of bloat. Customers no longer need the whole suite. They need the right tool. And if yours does not fit them precisely, someone will build one that does.
The answer is not more features. It is the right value, delivered to the right customers.
Start With Your Customers
Segment them ruthlessly, not by industry or company size, but by how they actually use your product to create value. Joe the Baker uses your platform to manage orders and stock. Anthony the Accountant uses it to close books and file returns. Johanna the Mortgage Broker uses it to track applications and chase lenders. These are not the same customer in different clothes. They are three different products waiting to be built.
This segmentation transforms you from a SaaS business that solves X, into a growth partner for Joe, Anthony, and Johanna. That is a complete shift in how you view your business. Not a service provider. A growth partner.
That is the real moat.
Spin a small, focused team around each segment. Not a traditional product squad that ships features against a roadmap while the customer is held at arm's length. Something fundamentally different.
Put a founder at the centre. Not a product manager. Not a team lead. A founder. Someone who owns the direction of the product, not just its delivery. Someone who stands in front of customers and takes responsibility for outcomes. Someone who knows their customers by name, understands their business deeply enough to see problems the customer has not yet articulated, and is accountable for growth, not just output.
The wider organisation becomes their support pillar. Infrastructure, legal, compliance, brand, existing customer relationships. These are the structural advantages the internal team carries that no external startup can replicate on day one. But the team operates with the autonomy to move fast, to decide, and to own the outcome.
AI agents give your team the ability to cluster tightly around each partner, with focused, value-based delivery. As the cost of code drops and time to market accelerates, these arrangements make more commercial sense than ever. And here is the upside you might not have considered: a smaller, more targeted product will attract customers who previously looked at your platform and walked away because it felt too big, too complex, too much. You were not losing them to a competitor. You were losing them to complexity. Unbundling fixes that.
The Rise of Micro-SaaS
Here is the uncomfortable truth: this model is not theoretical. It is already in production, and it is being built by people who are coming directly for your customers.
Right now, small teams, sometimes just two or three founders, are building focused, vertical software products aimed precisely at the customers big SaaS abandoned in its pursuit of scale. They are not building for everyone. They are building for Joe. For Anthony. For Johanna.
And they have something big SaaS does not: no legacy to protect and nothing to slow them down.
What makes micro-SaaS genuinely dangerous is not just focus. It is that AI has removed the primary barrier that kept small teams from competing. Development, marketing, customer research, outreach, support. AI agents carry the operational load that previously required entire departments. A team of three, properly augmented, can move with the velocity that once required a team of thirty. The economics of small now make sense at scale.
But the deeper advantage is not the technology. The technology is table stakes. The advantage is proximity.
A micro-SaaS founder building for mortgage brokers is in the group chat with their customers. They know when regulations change, when a workflow breaks down, when a single feature would save two hours a day. The product evolves because the relationship is real. The customer is not a user segment in an analytics dashboard. They are a partner in an ongoing conversation about their own business.
The customer who feels like a partner does not churn. They refer. They co-build. They become your distribution channel without ever being asked to.
AI gives you the capacity to compete. The relationship is what gives you the moat.
The Playbook Is the Same
Whether you are a big SaaS company spinning a vertical team or two founders building from scratch, the model is identical.
Get specific about who you serve. Put the customer at the centre, not the product. Give someone founder-level ownership of the outcome. Let AI carry the operational weight. Build a relationship that makes leaving feel like a genuine loss.
SaaS is not dying. The version of SaaS that tried to be everything to everyone is dying. What replaces it is smaller, sharper, and far more personal.
The question for every founder right now is not whether this shift is coming.
It is whether you move before someone else does it to you.