How do you build a competitive advantage when AI makes copying easy?

You build a competitive advantage today by becoming something there is less of. AI makes what you sell cheaper and easier to copy every month, and your competitors have the same tools you do. The edge moves to whatever stays scarce. If something that used to work is working less, your performance didn't change. The ground under you did.

The decisions a $10M company makes over the next 12 months matter more than they have in a long time. Sometimes one counterintuitive call that your competitors don't make is the entire difference.

So I want to walk you through seven shifts I'm watching in the market right now. None of them are finished, most have barely started, and every one of them is speeding up. For each one, I'll start with something most owners still believe, show you what's actually happening underneath it, and give you the move I'd make. Stick with me to the last one, because it goes against advice I gave people for years.

Why is it harder to grow a business when it's easier than ever to start one?

It's harder because starting is easy for everyone else too, and the number of companies chasing each customer is growing much faster than the number of customers. You've heard "it has never been easier to start and grow a business" a thousand times. The starting part is true. The growing part is the half nobody has really thought through.

Try something with me. Pull out your phone and count the apps on it. Maybe it's 50, maybe it's 80 if you're a little wild. Now ask yourself how many new ones you downloaded this year that nobody forced on you, like your bank's app. Was it more than two? For most people it's zero or one. Your phone, like almost everyone's, is a pretty static place.

Now look at the other side of the screen. It's been widely shown that the number of new apps released on the App Store doubled in the first half of 2026, something like 560,000 new apps in six months. Over the same stretch, downloads went up about 2%. You don't need to be an economist to see the problem. Supply doubled and demand barely moved.

Once you see that pattern of flat demand and exploding supply, you start seeing it everywhere. Back in 2019 the streaming platforms took in around 40,000 new tracks a day. Now it's around 150,000, and on some platforms close to half of those are fully AI generated. Americans filed well over five and a half million new business applications last year, a record and more than 60% above 2019.

You might think that's just digital stuff, cheap to throw up and meaningless. But it shows up in heavy industry too. You can't vibe code a car company, and yet China now has something like 129 electric car brands, of which maybe 10 or 11 actually make money, with a record 156 new models launching there in the back half of this year. When I was a kid, you could count the car brands on your fingers.

So here's what that means for you. The second a customer problem shows up anywhere, ten companies pop up to solve it. The tools get better and cheaper every few months, so next year it's twenty companies, and the year after that it's fifty. The number only goes one way.

Making money isn't complicated. It comes down to supply and demand, an idea everyone thinks they understand and very few act on. Raise demand and lower supply and you earn a lot. The trouble is that most companies respond to this pressure by pushing on demand alone: more ads, more leads, more content, more cold email. Picture a $10M company that watches five new competitors enter its category and responds by doubling ad spend. It feels like action, but demand for most things is fairly fixed, and when supply outpaces it, no ad is good enough to fix that.

The question I want you asking instead is about supply. How do we become something there is less of? How do we build or grow into something that gets scarcer while everything around us gets more abundant? Scarcity is what unpicks this whole riddle, and almost every other move in this article is a way to raise the scarcity of what you sell.

Why is AI making what you sell less valuable?

AI is making what you sell less valuable because anything that becomes easy, cheap and abundant loses its value, and AI's whole job is to make things easy, cheap and abundant. The AI optimists say "I can build anything now" as if that's great news for business. For the person selling the thing that just got easy, it's the opposite.

Think about water. Water is more valuable than anything you will ever sell. It literally keeps you alive. It also falls out of the sky, so nobody can charge much for it. Value doesn't come from how good or how important something is. It comes from two things together, desirability and scarcity. Take away the scarcity and the value goes with it, even when the thing is still just as desirable.

That is exactly what AI has spent the last three years doing. It takes things that were scarce and makes them abundant. Research, copywriting, analysis, design and software expertise were gold two or three years ago, and you could sell them for real money. AI has turned most of that into lead. Every new model knocks something else off the scarce pile and onto the abundant one, so the list keeps getting longer.

Most people understand this on paper. The problem is that their mental models haven't caught up. They're walking around with an imaginary price list in their heads that still says 2022, and they use it to decide where to spend their time, where to spend their money and what they think will help them win.

Think about the company blog that pulled in steady leads in 2022 because nobody else in your space could publish a thoughtful 2,000 word piece every week. Those posts weren't winning because they were brilliant. They were winning because they were rare. The moment every competitor could produce something similar in 30 seconds, the value vanished, even though the posts were exactly as good as before.

Content is just the tip of the iceberg. Picture an owner who builds an internal tool over a weekend with Claude, and it's better than the $400 a month software the company has been paying for. The first instinct is "we should sell this." But if it took 20 minutes to build, it takes your competitor 20 minutes too. Build it, use it, give it away to your customers if you like, which is exactly why I push companies to own their software instead of renting it. Just don't build a business on selling it.

You have to separate two ideas that people keep blending together, usefulness and commercial value. Water and air are incredibly useful and nearly worthless commercially. Gold has very little practical use and a lot of value, because it's scarce. If you have to pick, sell the gold.

You can watch this happen to big public companies. Chegg charged students about $15 a month for textbook answers because there was no free alternative that did the job. Then ChatGPT arrived.

Chegg's revenue fell 39% in a single year, from roughly $600 million to around $377 million, and the stock sits something like 99% off its peak. In software, around $285 billion was wiped off SaaS stocks in a single day this February because the things those companies sold suddenly got cheap. When there's real money on the line, investors update their mental models fast. Owners and CEOs are much slower.

Which parts of your business are gaining value and which are losing it?

Your product or service on its own, your technology, your IP, your cost structure and your expertise are losing value. Brand, distribution, proprietary data and scarce physical assets are gaining it. The first pile is getting easier for everyone, so it's worth a little less every month, though not at the same speed. The second pile was always hard and is staying hard, or getting harder while everything else gets easy.

So here's the first job. Go through everything you believe is valuable about your business, every form of leverage you think you have, and ask one question of each: is this still true? Is it still scarce? Then sort it into those two piles. It isn't an exact science, but it's close enough to show you where your attention is placed and what you're counting on to win.

Once you've sorted it, two questions follow. First, is our leverage growing or under threat? If your advantage sits on the falling pile, you're standing on a melting iceberg, and the only real question is how long until you're in the water with the penguins. Second, how do we build genuine leverage in one of the rising categories? You don't need all four. One or two defensible edges in places that are hard to win is enough, and that's where the next three years of your money, talent and attention should go.

Is AI making every company's strategy the same?

Yes, and it's a bigger problem than most owners realize, because your edge comes from being different, not from being better. When the pressure I just described hits, people do one of two things. They copy what the most successful companies are doing, or they ask ChatGPT what to do. Both pull you toward everyone else.

Have you noticed that every business guru online gives nearly identical advice? Sell the outcome. Niche down. Build an irresistible offer. Stack the value. Find the pain point. Fifteen years ago almost nobody talked like this. So did we all suddenly crack the code of business? Unfortunately, the truth is more boring. The algorithm happened.

The algorithm decides what information reaches you, and it doesn't decide by asking whether something is true or useful. It decides by popularity. Imagine three perfectly valid ways to run a sales process, and one of them happens to be snappier to explain in a 30 second clip. That one wins and the other two get buried. What we now call best practice is really most engaging practice, and everyone has been drinking from the same narrowing trough for a decade. That was before AI.

Now pour AI on top. AI doesn't go looking for the best answer. It goes looking for the most likely answer, because that's the only way it can work, and the most likely answer is by definition the average. Worse, it was trained on that same algorithm filtered internet you scroll every day. So you're getting a filter on top of a filter.

We now have a pretty good idea how far this goes. It's been shown that when you put two dozen commercial AI models up against more than a hundred people on the same creative problems, each AI's answers hold up fine on quality. But as a group, the AI answers look far more alike than the human answers do. They aren't thinking worse, they're thinking less differently. And telling the models to be more creative barely changes it.

That matters because leverage never came from the quality of your analysis. It comes from difference. Sometimes a worse answer wins, simply because it isn't what your competitors are doing. Raising quality while also raising convergence is a dangerous trade.

It doesn't much matter which model you use or how you prompt it. You and every competitor are being pulled toward the same strategies, the same websites, the same tone of voice, the same sentence structure and the same business model. And it gets worse from here, because as more of the internet is written by AI, the next generation of models trains on that, and the pull toward one version of the truth gets stronger.

There's a second effect owners don't want to hear. For most companies, AI isn't making them more money. It's been widely reported that about 95% of organizations see no measurable return from their generative AI projects. Part of that is because most companies use AI to answer the question "how do we do everything a little worse but a lot cheaper," which was never a revenue strategy. The bigger reason is that everyone gets the same gains at the same time. You picked up a lot of capability, and so did every competitor, so none of it turned into leverage.

What should you use AI for, and what should you keep for yourself?

Use AI for closed problems, the ones that have a right answer, and keep open problems, the ones about which path to take, for yourself and your team. Summarize this contract, check this forecast, build this model, draft this SOP: those are closed, and AI is excellent at them. Which market to enter, what to stop doing, how to position against a new competitor: those are open, and that's where you should give AI the least power over you.

We forget that not everything worth knowing is written on the internet. There's more knowledge outside it than inside it, including whatever sits uncodified in your head and your team's heads after years in your market. So don't overestimate what AI knows about your situation. And even if it did know all of it, the convergence problem would still pull its advice toward the average.

The second move is personal. Deconverge your own knowledge. If everybody is drinking from the same trough, whatever Instagram and LinkedIn happen to serve them that day, go drink from a different one. Read books, especially old ones and odd ones that never went viral. Talk to practitioners who get paid for results instead of views, and spend time with idiosyncratic people. They won't always be more right than what you see online, but they'll be different, and that difference is your advantage. Knowing peculiar things is just as valuable as knowing more things.

Why isn't your marketing working anymore?

Your marketing isn't working because attention is the one thing in the economy that's shrinking while everything competing for it multiplies. The belief is that good marketing saves you when all else fails: rank on Google, run the ads, send the emails, post the content, and leads and sales follow. The reality is that more and more of that is talking into the ether.

Your customers have the same 24 hours they had ten years ago, or a thousand years ago, and probably a lot less patience. Every brand, every creator, every news alert and every AI generated video is fighting for that same fixed amount of attention, and every new tool that makes content cheaper to produce makes the fight worse. You can't opt out of it either, because people can't buy what they never notice.

You can see it in the numbers. It's been widely reported that in the first months of this year, 68% of Google searches in the US ended without anyone clicking on anything, which is strange when you remember that the entire point of a search was to find something to click. When an AI Overview sat on top of the results, clicks on the number one result dropped by about 58%. So you can rank first and still be invisible.

The natural response is "fine, I'll pay for attention instead." Meta's average price per ad rose 14% last quarter, and 23% in the US and Canada. Email, which used to be the safe haven, now carries something like 392 billion messages a day, much of it garbage sent by very productive AI agents. Plenty of people are competing for the same eyeballs you want.

It's actually worse than more competition, because of what overwhelmed people do. They don't filter more carefully. They retreat. They fall back to the one newspaper, the two newsletters and the three creators they already trust, a tiny closed set of names. Think about yourself for a second. How many new sources did you find this year that you genuinely started trusting? That retreat is why attention doesn't just stay flat. It contracts.

The conclusion is hard to escape. In a world this abundant, whatever wins the attention war wins, and classic good marketing won't win it, because there's so much good marketing now that in effect there is no good marketing. The moment a marketing technique becomes easy to scale, it stops working.

What marketing still works when everyone has AI?

Marketing that is unfair, tedious or expensive still works, because those are the only kinds your competitors can't copy with a prompt. If it isn't difficult, it isn't going to work. The techniques where you'll find real ROI from here on are the ones most people avoid.

Unfair means using advantages you have that nobody else can copy. Here are nine to think through:

• Fame, inside your market or outside it

• A verifiable, impressive track record, the kind you can actually show

• Association with big names your buyers already respect

• Genuine insight of a kind no LLM can generate

• Costly signaling

• Real world relationships

• Showmanship

• Live demonstration

• Physical presence and things you do offline

If you can tick three of those, you probably have a reliable way to earn attention, precisely because all of them are hard.

Tedious means choosing the work everyone else can't be bothered with. Original research. Your own events. Fifty speaking slots in a year. Getting on the plane and working the trail. Handwritten letters to your top 200 accounts. In a world with no friction, adding friction is how you build a moat. The tedium is the point.

Expensive means the work where the ROI looks awful on the surface. Costly signaling is spending in a way so visible and so obviously expensive that the spend itself proves you're serious. Why does a company buy a full takeover of Times Square or a major train station? Because everyone knows what it costs, and they walk past thinking that company must be the real deal. After twenty years of everyone being told marketing could be made cheap, quick and easy, the companies that refuse that route will stand out.

Here's a test I'd run on every campaign before it launches. If this cost ten times more, would we still do it? Would we still be confident it works? If the answer is no, you're probably producing slop, and you need to go more tedious, more expensive and more difficult. That's the kind of marketing I mean when I talk about the marketing engineer role: building what's hard to copy, not more of what's easy to make.

Is your industry safe if nothing has changed in years?

No. A stable industry usually means the aggressive operators haven't shown up yet, and AI is about to send them. Plenty of owners will never read an article like this, because they're thinking "we have a great track record, we've done this for twenty years, we're safe." That's complacency, and the age of complacency is over.

To see why, admit something uncomfortable. For most businesses, mediocrity has always been more than enough. You could build a perfectly good company, sometimes a very big one, by doing a decent job in an industry where most players weren't trying that hard. Most industries run on an unspoken non-aggression pact. Every incumbent has its patch, it's a bit of a lifestyle business, everyone plays golf together, and everyone quietly agrees not to rock the boat.

Here's a strange question that explains why that pact is breaking. Why is there no Uber of electricians, no Netflix of school photos, no Airbnb of IT recruiting? It has nothing to do with those industries being special. They were simply too tedious and too low ROI for the most talented, most aggressive entrepreneurs to bother with. Those people looked at them, decided the bigger rewards were elsewhere, and walked away. Tedium was the moat that protected most industries.

AI makes tedium cheap. Building, distributing and running the boring parts gets easier every quarter. At the exact same moment, the opportunities in software are shrinking because the big AI companies are swallowing them. So where does all that aggressive talent go, the people who wanted to build a tech unicorn and now can't? It goes looking for something else to disrupt, and the obvious targets are the sleepy, underleveraged industries that have been coasting for decades. Maybe yours.

A frontier operator is a founder or executive who walks into a slow industry with a winner take all mindset, cutting edge tools, good taste, natural strategic instinct and no interest in work life balance. Compared to the average CEO in a cozy industry, they're sharks. When they arrive, they swarm.

You can already see the first wave. Last September, KKR agreed to buy A1 Garage Door Service for around $2 billion. Garage doors, about as sleepy as an industry gets. The founder started with a single van in 2007 and built one of the biggest garage door companies in America, mostly by running it at a level that industry had never seen. He didn't invent anything.

He showed up and tried much, much harder than everyone else. As the tools keep improving, the cost of walking into an industry like that and taking it over keeps falling, so expect a lot more of these stories.

So ask yourself an uncomfortable question. If the top players in your industry have been roughly the same for 10 or 15 years, what does that really tell you? It feels reassuring. It isn't. It doesn't mean you're special or the industry is especially hard. It means the ambitious people haven't decided to attack it yet. It's like playing high school basketball and deciding you don't need to train at Michael Jordan's level because he's in the NBA. That logic holds right up until Michael Jordan walks into your gym.

When these operators arrive, they consolidate, and the gap between winners and everyone else widens. It's been well documented that among the world's largest companies, the top 10% capture around 80% of all the economic profit, and that split has grown much more lopsided over the last 20 years as technology rewarded scale. The middle 60% make roughly nothing.

How do you become the frontier operator in your own market?

You become the frontier operator by attacking your own market before someone else does, and it starts with one question your leadership team can't lie their way through: what is the chance we triple this business in the next 18 months? I don't ask teams whether their strategy is good, because everyone talks themselves into a yes instantly. It's very hard to fool yourself about tripling.

If the honest answer is that there's almost no chance, then you're not being strategic or aggressive enough, you're not trying to create new value, and you're very exposed to whoever arrives and does. The question works because you can't triple a business by working harder or doing things a bit better. Effort only gets you to the ceiling of your current business model. Anything above that requires building a different business. Big ambition forces strategy. Small ambition lets you get away with tinkering, managing, turning the handle and waiting for someone to come along and eat your lunch.

The second exercise is a thought experiment. Picture a team of ravenous, well funded tech operators entering your industry tomorrow, if they haven't already. How would they go about it? Where would they beat you? Can you see it? Write it down, because you're writing your own playbook. In 2026 the only real defense is attack.

Is being better than your competitors still a strategy?

No. Being better at the same thing everyone else does mostly turns you into a commodity, especially now that anyone can copy and iterate faster than ever. It sounds logical, do what the industry does but at a higher level, and it's what most companies aim for. The real opportunities sit on the other side of the rules nobody in your category questions.

Every industry runs on a set of invisible rules: how you price, who the customer is, what the product has to include, which channel you sell through. Nobody decided these. They grew organically, and now they sit there like wallpaper. Everyone copies everyone else without noticing, until the rules become the water the whole industry swims in.

In a crowded market, those shared rules are exactly what lock everyone into the commodity fight. Follow the same rules as your competitors and you'll ask the same questions, get the same answers and end up in the same place.

You'll never escape an oversupplied market by doing what everyone does a little better, even if that were possible, because the market doesn't notice better. Plenty of better products have died to prove it. The market notices power and it notices difference. When every spot on the map looks taken, new space only shows up on the other side of a rule everyone follows.

My favorite example is the Savannah Bananas. Their category is baseball, which literally has a rulebook, but the rule every team really organizes itself around is the unwritten one: try to win ball games. That's where every team directs its energy and creativity. The Bananas asked what happens if we don't try to win ball games. What could we do that no other team ever does?

The answer was turning every game into the best fan experience in the country. The team isn't great at baseball, and people pack the stands anyway, because they were always in the business of selling tickets.

The results speak loudly. They sold out more than 50 cities in a season, including about 81,000 people in a single college football stadium, and their current tour runs 70 stops in major league stadiums where you have to enter a lottery just to get a ticket. No real baseball team can copy them, because copying them would mean dismantling the thing their own business is built on. That's what strategy is: a move your competitors are unable to copy.

Nintendo did the same thing with the Switch. To build a portable console that dominated party gaming, Nintendo gave up raw power, the one thing Sony and Microsoft will never sacrifice.

It works at small scale too. Picture a small design agency that notices every agency assumes it should be the guardian of the client's brand. It flips that rule and sells capability instead, teaching clients to run their own brand while the agency gets out of the way. That's an easy sell to most brands, and it's how a shop like that can double revenue in a year and lift its margin roughly tenfold.

How do you find the rules your industry never questions?

You find them by listing everything in your industry that's so fixed it has become wallpaper, then starting with the ones you're already worst at. A sacred cow is a rule your whole category follows without ever having decided to, like billing by the hour or selling through distributors. At first they're genuinely hard to see, so go category by category: the pricing model, the channel, who the customer is, what features are expected.

Start with the rules where you already have no advantage, because that's where you have the least to lose. Then ask the slightly psychopathic question: what if we were even worse at that? What if we didn't do it at all? What does that let us do that nobody else can? What if a law firm didn't bill by the hour? What if an advisory firm refused to run workshops? What if your product was deliberately worse at the exact thing every competitor fights over?

Most of the answers will look stupid at first. One of them won't be, or it will hand you a clue to something you never thought of. When you find it, it'll be so obvious you'll be a little embarrassed nobody noticed it before. That embarrassment is how you know you've found a defensible, high leverage winner.

Does a great product still sell itself?

No. For years I told people to build an outstanding product and let it sell itself, and I don't believe that anymore. I used to get sick of companies patching a weak, undifferentiated product with marketing. Don't put lipstick on a pig, I'd say, build the differentiation into the product and the rest will follow. But if anyone, relative to their industry, can build almost anything, iterate on it instantly and copy what works, product differences don't last.

There are exceptions. Some products get better the more people use them, like Instagram, and you can't copy a network effect. Some products are bound up with an asset nobody can replicate. If I want to build a hotel like the Ritz in London, I can't get that piece of land. And some businesses sell limited access, through licensing and the like. Most companies don't have any of these. Their products are hard to defend.

So I think we're heading into something close to a post-product world, and the faster the tools get, the sooner we'll arrive. If that's true, the order in which a business wins has flipped. It starts with attention, the scarcest thing your customers have. That attention goes one of two ways. It goes toward things that attract them, which covers most discretionary buying, or toward things that irritate them, which covers most non-discretionary buying. Think of a jacket you want versus a furnace that quits in January.

From there, the business has to hold that attention and turn it into something. If it started as attraction, you turn it into desire through brand, story, content and association. If it started as irritation, you turn it into relief by being sharply relevant to the problem and easy to reach, so the customer's reaction is "thank God, here's someone who can fix this."

Only at the very end of that chain does the product come in. By then, all it has to do is live up to the promise, because the battle was already won in the customer's head. How differentiated it is matters far less than you'd think.

This is why every celebrity and influencer is launching a drink, a skincare line or a protein bar. They've already won the top of the chain. The algorithm sends attention their way, and people buy wherever that attention lands. Nobody buys a celebrity's tonic water for its science backed claims. They buy it because of who's behind it.

So here's the move. You might be well advised, not always but often, to build your business out from attention rather than out from the product. Think of yourself as a media business with a product bolted on, not a product with media bolted on. Most companies put their best people, their money and their energy into product and hand attention to a junior marketing hire and an agency. Flip that. Attention is now the heart of the business, and it's where I spend most of my own work inside the companies I run marketing for.

That doesn't mean the product stops mattering. It can't be bad, and it still has to deliver fully. But it isn't that hard to make a solid product anymore, so stop pouring everything into making it 10% better when that 10% will be copied by someone in a quarter. Ask instead: what is our claim on people's attention? Where is our attention leverage? And what are we turning it into?

Look back across all seven of these and you'll notice they share one thing. Every move pulls the supply lever. You're trying to become scarcer by building leverage in the categories that are rising, deconverging your thinking, making your marketing unfair, tedious and expensive, attacking your own industry before a shark does, killing a sacred cow and building out from attention.

All of it is hard and still very unusual, and most companies will never do any of it, or they'll hear it three years from now when it's too late and copy whatever the algorithm calls best practice by then.

They'll skip it because it's hard and tedious, and they'll miss that the hardness is the entire reason it works. If you're still aggressive and still believe 2027 can be your best year, this is what to be working on right now.