What AI can’t copy

Hi Everyone,

A lot of what protected a business five years ago can now be copied by a good team with an AI subscription – product features, industry knowledge, even data. At the same time, a few advantages have become stronger than ever.

Today we're covering both lists, plus a simple test to sort your own advantages into weak and strong (plus a budget decision to make with the result).

Three advantages that became easy to copy

1. Product features: AI tools now can do most of the building work, whether that's code, content, or design, so a feature that took your team months to build can be copied in a few days at most. If your pitch is "we're the only ones who can do X," that claim won't hold for long.

2. Data anyone could collect: Companies like to say they have proprietary data. The question is whether a competitor could collect similar data: CRM records, support tickets, and usage logs usually don't pass that test. Data only protects you if others cannot get it.

3. The cost of leaving: Some products survive because moving to a competitor takes weeks of setup work. Your competitors know this, and AI assistants that handle the migration are becoming part of their sales pitch.

Four advantages that became stronger

1. Data no one else can get: Thomson Reuters built its legal AI assistant, CoCounsel, on more than a century of legal documents, many never published anywhere. In February, CoCounsel reached one million users, and the stock rose 11% in one day. A competitor with a better AI model still can’t get that archive.

2. Taking responsibility for the answer: In legal, tax, and audit work, AI can write the answer in seconds. Clients still need a firm that checks the answer and takes responsibility if it's wrong. That responsibility is what they pay for.

3. Running part of the customer's business: Think about what your customer would have to do to leave you. If they can cancel and move on in an afternoon, you're easy to replace. If leaving means retraining their team and risking that daily work stops, you're not.

4. Speed of your own team: A presentation company called Gamma rebuilt its whole product around AI in three months, and now keeps winning customers even though Microsoft and Google offer AI slide features for free. Fast teams get more out of AI than slow teams, so the difference between them keeps growing.

Test your own advantages

Write down the top three advantages your company claims in its sales deck or investor story.

For each one, answer this: Could a good team with AI tools copy this within three months?

If yes, the advantage is weak. If something specific stops them, like an archive, a license, a long-term contract, or a process you run for the customer, the advantage is strong.

The same test works for services companies.

Knowledge alone is weak, because AI models are learning it. Knowledge plus responsibility is strong, because clients pay for someone who knows their situation and takes responsibility if something goes wrong.

What to do with the results

Companies leading on AI move serious money and people to new priorities during the year, instead of waiting for the next annual plan. In McKinsey's research, they were 3x more likely to do this than slow companies.

So finish the test with one change. Find one weak advantage your company still spends money on, and one strong advantage that gets less money than it deserves. At your next planning meeting, propose taking budget from the first and giving it to the second.

Go deeper

👉 McKinsey: Why AI first movers win through speed and resource allocation — read this before your next budget cycle; the five actions at the end pair well with today's test.

👉 BCG: The Widening AI Value Gap — use it to see what the top 5% of AI adopters do differently, with revenue and margin numbers to back it up.

👉 Steven Sinofsky (a16z): The Four Stages of Disruption — read this to place your own market on the disruption timeline before a competitor does it for you.

👉 Gamma: How we built a $100M business differently — the founder's own account of reaching $100M with 50 people and almost no outside money.

Coming up on Monday

On Monday, you'll learn how to decide whether to counter an outside offer or let a key person go.

Have a great weekend!