Every business owner who discovers AI search sooner or later discovers Wikipedia, and the first idea they have is to write themselves an article. Unfortunately, Wikipedia’s editors have watched people try that move for a quarter century, and they built the whole place to stop it. They’re right to. I still find it maddening.
Think of a Wikipedia article as a letter of recommendation. A letter like that counts because somebody else signed it.
Would you trust a reference letter the candidate wrote himself? Neither would anyone else. Once a reader suspects you wrote your own, the letter means nothing, and Wikipedia’s volunteer editors are the most suspicious readers you’ll ever meet. Every rule below exists to answer one question: did somebody other than you decide you were worth writing about? Without that, the article dies. Good prose won’t save it, and neither will the hours you spent on it.
Why can’t you write your own Wikipedia article?
Wikipedia’s own rules tell you not to. Its conflict of interest guideline says COI editing is strongly discouraged, and it tells people to refrain from creating articles about themselves, or anyone they know, unless they go through Articles for Creation. That’s a queue where a volunteer reviewer reads your draft before it goes live. If somebody’s paying you to edit, the Wikimedia Foundation’s terms of use require you to disclose who’s paying and on whose behalf.
My friend Joe Rockey asked me on a call this spring whether it made sense to write his own Wikipedia article. You can’t write an article about yourself. You can write a Wikipedia article about some subject, but you can’t promote yourself, and you can’t make it about you or your business. They’ll delete it.
So Joe asked whether he could write one about me.
He could, and it’d probably get deleted, because I don’t yet have the coverage the rules demand.
A friend’s keyboard doesn’t change the evidence.
What does Wikipedia notability require for a person?
Significant coverage in reliable sources that are independent of you. That’s the general notability guideline almost word for word, and every word in it does work.
Significant coverage means a source addresses you directly and in detail. A passing mention in a list of speakers doesn’t qualify, and neither does a quote buried in somebody else’s story. Reliable means the source has editorial integrity: an editor, a fact checker, somebody whose job is to be wrong in public as rarely as possible.
Independent rules out anything you or your people produced. The guideline names advertising, press releases, autobiographies and your own website as sources that don’t count, and it says flat out that self-promotion and paid material aren’t valid routes to an encyclopedia article.
Then comes the word that catches people. Coverage that meets the test creates a presumption that you merit an article. Editors argue plenty of presumed subjects right back out.
My shorthand for all of it is other-people-done. For a person, you need interviews and profiles in big magazines and on sites like Forbes, a lot of public material, and none of it self-done. My rough count is about a dozen real attributions, from places like CNN, before I’d try. Wikipedia publishes no number, so treat the dozen as my working bar. A biography propped up by three thin sources gets torn apart in a deletion discussion.
Why I dropped the Wikipedia editor route
Last spring I had a plan, and it was a big one. Becoming a Wikipedia editor was one of the big items on a long list of things I was doing to make AI search engines know who I am. You can’t write about yourself, but you can become an editor, gain trust, and then another editor can write the article about you.
The plan was to go in every day and edit a couple of Wikipedia articles for six months. Every edit is marked. Wikipedia knows who changed what, when they changed it, and where the information came from, and senior editors come through and review the changes.
You have to be very careful never to promote yourself in any of it. The example I gave Joe was Ulysses S. Grant’s birthday. If I change it because the article has it wrong, I’d better cite this book and that book and a third one that all back me up, because Wikipedia runs on citations. It’s a pain in the ass. I figured on three or four edits a week, no more.
Then I added it up. I put the whole Wikipedia effort at around two years, and the deeper I dug, the more work it looked like, for a payoff years out. I’m 65. I don’t know that I want to spend two years editing Wikipedia articles for an article nobody can promise me.
By July I’d made the call. When Joe asked whether I’d dropped the Wikipedia angle, I told him I have to build from the other side: interviews with real outlets, real attributions, about a dozen of them. They also want months and months of editing before they’ll maybe let you near anything about yourself, and like Reddit, it’s too much work.
I know why they do it, since spammers would bury the place otherwise. I don’t have the time.
How long does it take to become a trusted Wikipedia editor?
The technical bar is low and the trust bar has no number. On English Wikipedia an account becomes autoconfirmed once it’s four days old and has made ten edits, and since 2018 only autoconfirmed accounts can create a new article directly. Anyone can clear that in a week.
Trust is different, and nobody hands you a badge for it. Reviewers judge you from your edit history. That history stays public forever, every change tied to your username with a timestamp and a diff anyone can read. An account that edits a dozen unrelated articles and then shows up to polish a biography of a ghostwriter in Florida looks exactly like what it is, and editors will check your contributions page before they read your draft.
A year of spotless editing makes you a trusted editor. It does nothing to make you notable.
Notability lives outside Wikipedia, in newspapers, magazines and books written by other people, and no amount of work inside the site can manufacture it. Most people’s fantasy dies right there. The apprenticeship gets you a reputation as a volunteer and leaves you exactly as far from a biography as you were on day one.
Why does Wikipedia matter so much for AI search?
The machines read it. Wikipedia, as I told Joe on that spring call, is the number one input to AI and Reddit is number two, with sister sites like Wikidata and Wikiquote feeding in alongside. When an answer engine decides who you are, it leans hard on sources like these, and a person with a Wikipedia article starts every query with an advantage you can’t buy.
I don’t think much of either site as a source. Reddit and Wikipedia are both edited by people with their own biases, and AI writes at the lowest common denominator partly because it eats so much of them. That irritates the hell out of me.
My opinion of the source doesn’t change what the machines read.
A Wikipedia article is a letter of recommendation, and nobody believes the one you wrote yourself. – Richard LoweShare on X
Don’t bet everything on it, though. What you learn about AI search this year won’t hold next year. Quora used to be one of the big inputs, and then it filled up with spam and dropped out, as I wrote in Quora Sucks. Reddit or Wikipedia could slide the same way.
Stay nimble, keep an eye on what the engines cite, and spread yourself across several independent sources so no single gatekeeper decides whether you exist. The longer case is in how AI picks who it recommends, and the Wikipedia wall is the strictest version of the same test.
The people who sell Wikipedia pages
Search for help getting on Wikipedia and you’ll find agencies that promise to put you there for a fee. Be very careful. At the end of August 2015 Wikipedia’s volunteers uncovered a ring they called Orangemoody, blocked 381 accounts for undisclosed paid editing, and deleted about 210 articles written to promote businesses, entrepreneurs and artists. Those clients paid for pages that vanished, and their names ended up attached to a scandal in the bargain.
That makes me furious on behalf of the buyers.
A small business owner who hears that Wikipedia matters for AI does what any reasonable person does and hires somebody who claims to know the system. The seller knows the article will be deleted the first time a real editor looks at the sources, and takes the money anyway. Anybody guaranteeing you a Wikipedia page is selling you a deletion notice with a delay on it.
If you do hire help with Wikipedia, insist the person discloses the paid relationship on their user page and on the draft’s talk page, goes through Articles for Creation, and builds the draft only from independent coverage you already have. A writer who won’t agree to those conditions plans to break the rules with your name on the result.
How do you earn the press coverage Wikipedia wants?
One story at a time, starting closer to home than you’d think. A reporter at a regional paper writing about you in her own words counts for more than a dozen mentions you arranged yourself, and local outlets need stories every single day. The playbook for getting local media to cover your book works the same way here: your book is the credential that gets you in the door, and the story is what the reporter wants.
Answer reporters’ queries. Services built on the old Help a Reporter Out model send you journalists’ requests for expert sources, and a quote in a reputable outlet is a start.
One quote won’t get you there. You want the profile, the feature, the interview where the reporter spends paragraphs explaining who you are.
Pay attention to whose voice carries the piece. Many Wikipedia editors discount the parts of an interview where you’re talking about yourself, since that’s you describing you with a microphone in front of you. The reporter’s framing, the background research and the outside opinions in the article are the independent part. A trade magazine profile written by its own staff counts. A sponsored post with the same words doesn’t, because you paid for it.
I send out press releases. They don’t count toward Wikipedia, and I don’t pretend they do.
Each release is one more place where a reporter or an answer engine could notice my name and smell a story. Keep a running list of every independent piece about you, with the publication, the date, the link and a note on how deep the coverage goes. When that list holds a dozen real entries, you have something an editor can work with. Until then you have a hope.
What should you fix while the press coverage builds?
Everything you control, starting with your own website. Answer engines check what they find about you against your site, and if your site contradicts the other sources, you’ve got a problem. Mine is about twelve years old, so of course the numbers had floated, one page saying one count and another page saying something else. I’ve been fixing all of that, page by page, and adding structured schema so the machines can read who I am without guessing.
Wikidata is a separate project from Wikipedia, with its own rules. An item there has to describe a clearly identifiable entity using serious, publicly available references, and it holds facts in a form machines read directly. I set up my own Wikidata entries. A Wikidata item won’t stand in for a Wikipedia article, but it’s one more independent record that agrees with your site, and it’s within reach today.
Write the book. A properly published book puts your name into library catalogs and bibliographic databases you don’t control, all of them agreeing, and the full argument is in why a book is the strongest AEO asset you can build. Reviewers and journalists cover books. Nobody covers a services page.
My own guide to how answer engines decide who gets named is The Day Your Website Died.
Should a business owner become a Wikipedia editor anyway?
Only if you love the work for its own sake. Plenty of people edit Wikipedia because they care about Civil War history or obscure films, and the encyclopedia needs every one of them. Edit what you know, cite every change, and never touch an article about yourself, your company or your competitors.
As a marketing tactic, forget it. If you’ve got the time, you’ve got the time, and most business owners don’t. Two years of careful volunteer edits buys you a good reputation among editors and nothing else. The same two years spent earning real coverage moves you toward the article and helps you with every other source an answer engine reads.
Why would you spend two years on the slow road to nowhere?
The wall exists to keep out the spammers and the page sellers, and anybody in a hurry hits it too. I hate how long the slow route takes.
I hate the shortcuts more, because every one of them ends with your name on a deleted page and somebody else’s invoice paid. Go get the letters of recommendation written by people with no reason to flatter you, one reporter at a time, and let Wikipedia find you when there’s enough on the record to write about. The AI and Writing Hub covers the rest of what answer engines look for, and my AI services can help you build the record they read.
