What cannot be copied when all content sounds the same
Sameness did not start with AI
For about a decade, publishing an article ran on one method. Look at what ranks for the question. Read the top ten results. Cover the same ground more thoroughly, with cleaner formatting and a better table of contents.
The method worked for years. It also guaranteed convergence. Every publisher was studying the same page and writing toward the same target.
Then the target moved. Once a search result held ten thorough articles, thoroughness stopped being a reason to pick any of them.
Tactics decay once everybody copies them. This one decayed slowly, and then all at once, because the copying got automated.
About half of new English-language articles are now classified as primarily AI-generated. That share has held near half for five quarters. Producing an adequate article on any subject is now free and instant. So adequacy stopped separating anybody from anybody.
The flood is losing, which is the useful part
Volume did not turn into visibility. One study looked at search results and answer engines together. It found that 86% of articles ranking in Google were written by humans. Of the articles cited by ChatGPT and Perplexity, 82% were too.
One caveat belongs with that finding, and the researchers state it themselves. They did not test AI drafts that were heavily edited by a person. They suspect that combination performs well. Detectors also miscount in both directions.
So the finding is narrower than the headline. Writing produced with a machine is not what loses. Writing with nothing in it is what loses.
Readers notice the pattern before they can name it. What they are detecting is genericness rather than machinery.
Four kinds of material a competitor cannot copy
Start with numbers from inside your business. What you charge. How long the work takes from signature to delivery. How often the problem you warn clients about shows up, out of your last 20 engagements.
A model can estimate those numbers for your industry. It cannot know yours. Neither can the competitor reading your site.
Second, positions that cost you something. Name the popular tactic in your field that you decline, and give the reason. Name the advice you give that loses you a certain kind of client.
A position with no cost attached is a platitude. Platitudes are the raw material of the flood.
Third, cases with the details left in. The dates. The constraint you worked around. The thing that went wrong in month three, and what you changed afterward.
Smoothed stories are interchangeable, because a smoothed story is a parable and anybody can write a parable. A rough one is evidence.
Fourth, judgment at the edges. Every field has a textbook answer, and the textbook answer is the most reproduced text in the world. Your value sits in knowing where it stops applying. Say where, and say what you do instead.
None of that material is published anywhere. Getting it out of your head and onto a page is a problem of its own. It is also the reason so much of this material never surfaces.
Specificity has a price, and the price is why people skip it
A number can be checked, and somebody will check it. A dated fact ages, and you will have to update it. A position loses you the readers who hold the other one. Naming your price invites a comparison you do not control.
Generic writing carries none of those costs. It is safe writing. Safety is the reason for sameness that the advice columns never name.
Here are both sides of that choice. A reader who meets a specific position remembers where they read it. They quote it in a meeting, and some of them send the message that starts a conversation.
A reader who meets the safe version finishes it and agrees with it. Then they forget who wrote it. They hire the person whose article they could still name on Thursday. You never learn that the comparison happened.
The check I run before anything ships
Take a sentence you published. Put a competitor's name in front of it. If the sentence stays true, it belongs to your category rather than to you.
I run that check on my own pages. It also sits in the guides I write for clients. Most opening paragraphs fail it.
Failing is normal, and the repair is small. Replace the claim anyone in your field could make with the version only you can support. That usually means adding one number, one date, or one refusal.
"We take a personalized approach to every client" survives the swap. Any competitor could run it word for word. "Your launch date depends on your calendar, not mine" does not survive it. That line is only true of a firm whose own work is already done.
What specificity will not do
Specificity will not distribute your work. Uncopyable and unread is a common outcome. Publishing and distribution stay separate jobs, with separate budgets and separate skills.
It also will not carry one piece on its own. Specifics compound across a body of work. A single sharp article changes very little by itself.
Confidentiality limits some of it, particularly in law, medicine, and any engagement under an agreement. The pattern still publishes when the case cannot. An anonymized version with the number left intact keeps most of the value. Inventing a detail to fill the gap costs you the whole benefit.
One more limit, and this one is my reading rather than a finding. Generic content still performs in some categories, particularly the motivational lanes that were formulaic long before any of this. Where credibility is the product, it costs more than it returns.
Where your material sits
For most experts, the material is not scarce. It is unpublished.
It sits in the work you did last month. It sits in the question a client asked on Tuesday. It sits in the figure you know without looking it up. Most of it sits in the position you keep editing out of your drafts because it might annoy somebody.
Take one piece you published this year. Find the paragraph any competitor could have run without changing a word, which will not take long. Replace it with the paragraph only you could have written.
That edit, repeated, is the whole strategy.
