Not Using AI? That Might Be Fine — But Only If You Have a Real Moat

I've written several posts about why businesses should adopt AI.

Today I want to say something different.

Not using AI might actually be fine.

Consider this: Hermès reported full-year 2024 results with revenue growth of 15%, reaching €15.2 billion. Operating margin hit 40.5% — an exceptional figure in any industry. That same year, the global luxury market fell roughly 2%. LVMH and Kering were both in correction. Hermès widened its lead into the headwind.

The brand runs almost no AI production. Nearly every bag is handmade by artisans. They deliberately don't scale, deliberately stay scarce.

Also in 2024, Goldman Sachs research analyst Jim Covello published a blunt report: "Gen AI: Too Much Spend, Too Little Benefit?" The data he cited showed that a large share of AI implementations couldn't demonstrate measurable returns. BCG confirmed the same year that 74% of companies attempting AI still struggled to show tangible results.

So yes — companies that thrive without AI genuinely exist. And some companies that bought AI, deployed AI, still saw nothing move.

But put those two facts together, and they don't add up to "AI doesn't work."

The more accurate read is: not every business is starting from the same competitive position.

The businesses that are "fine without AI" aren't confident because AI doesn't matter. They're confident because of something you might not have thought carefully about — a moat.

Warren Buffett's favorite concept: an economic moat is a structural advantage that makes it hard for competitors to attack you.

Which raises the question: is your moat as deep as you think?

Are there really businesses winning without AI?

Yes — and not just Hermès.

In 2025, Italian brand Tod's launched an "Artisanal Intelligence" theme for its spring/summer collection, positioning its handcrafted process as a brand statement against the AI era. Prada's Miuccia described her Fall/Winter 2025 collection as "a quiet revolt against the first full season of AI design."

These brands aren't avoiding AI. They're actively managing "handmade" as a scarce asset. Choosing not to use AI is strategic, because using it would destroy the moat.

Beyond luxury, there's a much larger and more everyday group: businesses built on purely physical work.

Plumbers, construction workers, caregivers, physical therapists. The constraint isn't that AI "isn't good enough yet" — it's that AI cannot physically show up. If a pipe bursts, someone has to be there. U.S. Bureau of Labor Statistics 2025 data shows these roles are not shrinking due to AI — overall demand is growing steadily.

But all these cases share one thing in common.

It's not that they "chose not to use AI." It's that they already have a wall that keeps competitive pressure out.

What does a real "AI-proof moat" actually look like?

Morningstar systematically assessed 132 companies' competitive moats between 2024 and 2025, specifically examining which structural advantages AI disrupts most. Their conclusion was clear: the impact depends entirely on your moat type.

Across the research, five types genuinely hold up under AI competitive pressure:

1. Scarcity-based craft and brand

What Hermès customers are buying isn't "a bag that holds things." It's 185 years of accumulated artisanal reputation and the promise of never scaling. If Hermès started using AI to mass produce, the moat would disappear. They're not choosing the harder path — they're choosing to keep the moat intact.

2. Heavy regulatory protection

Core decisions in healthcare, law, and finance still legally require qualified humans to sign off. In these sectors, AI adoption isn't blocked by capability — it's blocked by compliance. Regulation creates structural barriers AI can't simply route around.

3. Purely physical work

Your service requires a human body to show up, with hands, eyes, and real-time physical judgment. However capable AI gets, it cannot actually hold a wrench, apply massage therapy, or make split-second calls on a construction site.

4. Relationship as product

A therapist, a deeply trusted family financial advisor, certain healing practitioners — what the client is paying for isn't information. It's the connection and trust with that specific person. If another stranger, or an AI, could substitute, you're not in this category.

5. Geographic monopoly

The only clinic in a remote mountain area, the only mechanic on an island — no competitors means no efficiency race, and no scenario where "a rival using AI overtakes you."

A 2024 MIT Sloan Management Review field experiment is worth citing here. Researchers gave half of a group of small business owners in Kenya access to GPT-4 as an advisor; the other half received nothing. The result: AI gave the biggest lift to entrepreneurs who already had strong judgment. Those with weaker judgment were sometimes actively harmed — led astray by AI advice they couldn't critically evaluate.

The researchers' conclusion: AI helps the best and may hurt the rest.

Applied to moats: if you're already the most irreplaceable player in your market, AI's marginal value to you is limited. But if you're in the competitive middle, AI is accelerating your rivals. Staying still is falling behind.

But do most SMBs actually have a moat?

That's what this post is really asking.

BCG's September 2025 report found AI leaders achieve double the revenue growth and 40% greater cost savings versus laggards. PwC's 2026 AI study confirmed that 74% of AI's economic value flows to the first 20% of companies to act.

Those numbers are usually cited to show "how effective AI is."

But I think the more important implication is: competitive advantage is concentrating fast. And it's concentrating first in the hands of those who were already strong.

Where does the other 80% go?

Some of it goes to the five moat types above — businesses that were never in this fight to begin with.

But a larger portion goes to businesses that believed they had a moat, when really they just hadn't felt the pressure yet.

"My clients have been with me for years — we have real relationships."

"I've been in this industry for twenty years. A newcomer can't catch up."

"Our market is small enough that competition isn't that fierce."

These feel like moats. They might not be.

Long-term client relationships can be eroded by competitors who are cheaper, faster, or simply easier to reach. Twenty years of experience is real — but if a competitor uses AI to compress learning time from ten years to two, the gap closes. A small local market insulates you from local competitors, but your clients' options aren't limited to your geography.

OECD's 2025 report found Japan's SME generative AI adoption rate is 23.5% — the lowest among all surveyed countries. Taiwan and Japan's small businesses are still largely watching from the sidelines.

That doesn't mean they're losing right now.

It means a gap is accumulating. And gaps accumulate until they suddenly become visible.

So how do you know which camp you're in?

Three questions. You can ask yourself today.

1. Do your clients choose you for who you are, or for what you do?

If a client could switch to another provider doing the same thing, would they hesitate at all? If the answer is no, your moat isn't the relationship — it's the switching cost. And cost-based moats are the easiest for efficient competitors to break through.

2. How long would it take a competitor to catch up to you?

A real moat means competitors need years to replicate what you have. If the answer is "within six months," what you have is a time advantage, not a moat.

3. How much of your core work is "roughly the same every time, with a predictable process"?

The higher the repetition and the more codified the process, the easier it is for a competitor to accelerate it with AI. If your competitive edge comes mainly from "doing it faster and cheaper than others" — AI is already eroding that edge, whether you use it or not.

Back to the original question: does it matter if you don't use AI?

If you're a multi-generational craft workshop, the only provider in a remote area, or a business built on irreplaceable personal trust — you genuinely don't need to rush.

But if you're in the middle — delivering knowledge-based services, process-based services, or anything that can be delivered remotely — then "not rushing" is letting the gap grow.

The point isn't whether you use AI.

The point is whether you clearly know where your moat is, and whether it's actually deep.

If you're not sure, start with one concrete question: in your business, which repeated task is hardest to hand off to a new person? The answer won't be in any AI tool. It'll be in how clearly you understand your own work. That clarity is the foundation of any moat worth having.

Not sure if your moat is real?

No software to sell, no package to pitch. Just a conversation about your situation and where your competitive edge actually comes from.

Talk to Barry

References

Back to Blog