How the Scale Game is Changing

Build it to scale. Reduce friction. Make it replicable at a low marginal cost. Then win! That was the game plan for any company lucky enough to find "product-market fit" (a solution to a genuine pain point).

AI is the apogee of that playbook.

AI will scale everything that can be scaled. Customer service. Content creation. Data analysis. Workflow orchestration. Legal review. Market research. Code generation. The entire stack of what I call "the first layer of friction". Monetizing services that address this layer is at the core of most business models.

So, AI should be a victory lap. All businesses ever wanted was frictionless scale, and we're getting it.

But there is a problem. Everyone is getting it at the same time.

And when something becomes a commodity, it stops being valuable. This is "the great inversion" I am predicting.

The entire venture capital model for the last 20 years has revolved around funding services or products that scale exponentially. Software scales. Marketplace scales. Network scales. Build it once, with your secret technical sauce (your IP or your moat), then replicate it liberally at minimal cost.

That was the entire game. AI will now change that game. Because anything that can scale can be scaled more efficiently with AI.

And what can scale? It's anything that is "decomposable" into a combination of actions that can be automated. It's anything, technical or not, that can be codified. Anything that can be captured or described with data. Anything that I will call "mechanical intelligence."

So, the question for all businesses becomes:

Is my IP, my competitive moat, decomposable into discrete items that can be orchestrated and automated by AI?

Most business executives ask this question, but not for the most important reason. They ask it because they want to "agentize" their company (see part 2 of this series). They ask it because they think that once "agentized", they will be more efficient, more competitive, and more attractive to customers. It's the current rush and pressure all of them feel in their bones.

Here is the issue. For sure, some companies will "agentize" better than others. But eventually, agentic efficiency will become a commodity.

It will be available to anyone with an API key and a credit card. And if your competitive advantage is a playbook that can be "agentized" then you just handed the blueprint to everyone. And someone with bigger pockets, with a better token budget, will beat you at the game.

As I said in part 2 of this series:

The supermarket will always beat the farmer's market on every mechanical dimension. Speed. Cost. Scale. Consistency. The gap used to be manageable. With AI, it becomes infinite.

So, what's left?

The farmer's market doesn't compete with the supermarket. It renders the comparison irrelevant.

It operates in a different value universe. One built on recognition, relationship, context, judgment that can't be written down because it only exists in the interaction itself. The vendor who remembers you from last week. Who set aside the good tomatoes. Who asks about the dinner party.

That's not a workflow. That's not scalable. That's the point.

Value is migrating to what doesn't scale.

The "uncodifiable" knowledge that lives in relationships, not systems. The judgment calls that require context you can't database. The human intelligence that comes from being present, not processing data. The trust that gets built over time in ways no CRM can replicate.

This will be a paradigm shift for an entire generation of business leaders and investors – raised with a "growth hacking" and "10X" mindset. Leaders and investors for whom small was unfundable, human-powered was a bottleneck, and efficiency was the holy grail.

AI is inverting that entire mental model.

The bottleneck is now the advantage. The inefficiency is the moat. The thing that doesn't scale - the thing you thought you had to fix - is the only thing left that cannot be commoditized by a better-capitalized competitor with more compute.

So, here is the more important question every SMB should be asking right now:

What do we do that is valuable because it doesn't scale? What part of our business only works because it's small, manual, relational, and human? What would we lose if we automated it?

Because that's your moat. Not your tech stack. Not your workflows. Not your efficiency gains.

The small and unscalable things that AI can't do better might be the only things worth keeping.

——————

by Eric Raza | March 22, 2026

This is part three of The Unscalable Advantage - a series on what survival looks like for SMBs in the age of AI.

The next piece reveals why this inversion isn't random. There's a pattern to what AI can and can't do. And understanding that pattern might be the difference between surviving and disappearing.