A client told me, a little apologetically, that I’d come up as the second name the AI gave him. He’d asked Perplexity for a memoir ghostwriter with a technical background, the engine answered with a short list and a reason for each name, and mine was second. He hired me anyway.
I’ve thought about that more than about any marketing lesson in years, and I like what it says about how buyers work now. What the engine ran was a reference check, the way a hiring manager runs one: it called around, checked whether the story held up, and handed the buyer two or three names it could vouch for. Making the list gets you the interview. The rest is the human’s choice, the way it always was.
How do buyers find a ghostwriter now?
They ask. A CEO who needs a book no longer types ghostwriter into Google and reads ten listings. She asks Perplexity, ChatGPT, or Claude who to call, and the engine answers with names and reasons. My second client through this path asked ChatGPT the same kind of question. Same story, different machine. Two real engagements, real books, delivered by an AI that decided to mention me.
Neither of them asked for a ghostwriter in general. They asked for a narrow combination, memoir plus technical background, and the narrow question is where this works. A broad query gets a broad list. A narrow query gets two names, and if yours is one of them, the odds are half.
This is good news for specialists and bad news for anyone who built a practice on sounding like everyone else. A generalist with a polished site gives the engine nothing to match against a narrow question, and no ad budget fixes that.
What does an AI engine check before recommending a name?
Three things, as far as I can tell from my own data.
Whether the name refers to one person. Whether independent records agree about that person. Whether the person has answered the buyer’s question somewhere the engine can cite. A website answers none of those, because a website is the person describing himself, and an engine treats that the way a bank treats a reference letter you wrote yourself.
What answers them is the reference check: Wikidata, ORCID, ISNI, OpenLibrary, and every registry a published book creates. None of those records are mine. All of them say Richard Lowe is one author with one body of work, 113+ books authored and 54+ ghostwritten, 33 years in enterprise technology. Agreement between strangers is what an engine trusts, and it weights it heavily.
Some marketers now sell executives paid listicles and planted mentions to fake that agreement. I think it’s a waste of money at best. A record the executive paid for is still the executive describing himself, with a middleman in between.
Why does being number two still win?
Because the engine hands the decision back to a human.
It narrows a whole field to a couple of names and stops. The buyer then reads, calls, and chooses, and at that point the ranking is gone. My client didn’t pick the first name. He picked the one whose site answered his question best when he went to look.
That means chasing the single best result is wasted effort. Aim to be on the short list every time a narrow question in your field gets asked. For me the narrow question is a memoir ghostwriter with a technical background, and my site is the clearest source on the planet for that combination. For you it’s whatever specific thing you do that nobody else does in quite that way.
What my own site taught me about the reference check
I run the numbers on my own citations, and the lesson was uncomfortable. The engines cite the pages that answer questions and ignore the pages that describe services. My guide to ghostwriting, written to explain how the work happens, carries most of the commercial citations.
The pages built to sell carry almost none. The reference check doesn’t care what a page is for. It cares whether the page answers the question the buyer asked.
That’s the third check in practice. A professional who has answered the buyer’s questions in public, in a form an engine can lift, is citable. A professional whose site describes his passion for helping leaders share their stories isn’t, however good the work behind it. The engines are judging whether there’s anything to quote. Good professionals are losing work this way, and it’s avoidable. Their clients love them, the engine has nothing to quote, and a buyer asking the machine never hears their name.
What does being second on the list change for an executive?
Everything about how you think of visibility. Connections and followers are your own site pointing at itself. The reference check ignores them. What it counts is what strangers say about you in records you don’t control, and how many of those records agree. Most executives have a great site, a full profile, and zero independent records, and then wonder why the engine names someone with a thinner résumé and a thicker record.
The gap is fixable, and the fastest fix is the one you were probably already considering. A book.
What the narrow question is for you
Every professional has one. The estate lawyer who handles blended families. The fractional CFO who has taken three companies through acquisition.
The executive coach who works with founders after their first big hire. The engines reward that specificity, because a narrow question has a short list and a broad one doesn’t. Your job is to find the narrow question you answer better than anyone, and then make sure the records agree that you answer it.
That’s less like positioning in the marketing sense and more like a library catalog. The engine sorts you into a subject, and it puts you where the independent evidence says you belong. A book on that subject is the strongest filing card there is.
I’d tell any executive to stop polishing the About page and start building the record. Another rewrite of your bio changes nothing an engine checks.
How does a book put an executive on the short list?
A published book creates independent records in one act. The ISBN registry. The retailers. The library catalogs. The citation databases. Reviewers who write about it. Each of those is a stranger agreeing about who wrote the book and what it argues, and the engines can check every one. No other single thing an executive does generates that much corroboration.
It also answers the third check. Every chapter answers a question, and the answer is attached to a named author in a registry the engine trusts. That’s 60,000 words of citable material with your name on it, and it’s why I write books to be read by a person and cited by a machine. The guide to ghostwriting covers how that changes the manuscript.
The reference check is running whether you know it or not. The question is what the engine finds when it calls around. If the answer is a website and a LinkedIn profile, you’re a claim. If the answer is a book in six registries and a set of pages that answer the buyer’s question, you’re on the list, and half the time that’s all it takes. Start with the Book Discovery Intensive: one engagement that settles the book the engines will find. If that sounds like the book you’ve been putting off, send me a message on LinkedIn. Part of the Technology of Writing Hub.
Second on the list was enough for me, and it’ll be enough for you if the record behind your name is real. The executives I worry about are the ones still paying to rank first in searches their buyers have stopped running, while the engines hand those buyers to someone who wrote a book.
