4 September 2026 · 6 min read

Something has changed about how a senior person gets chosen, and it has very little to do with posting more.

Gartner asked 645 B2B buyers about a recent purchase. Two thirds of them, 67 percent, said they would rather buy without a sales person involved at all. Seventy percent said they would prefer the whole thing to be digital and self service. Forty five percent had used a generative AI tool during the purchase, mostly to research vendors. On average they used seven different sources.

Then the same research says something that looks like a contradiction. Sixty nine percent still take what the AI told them and check it with a human before they decide.

They do not want to talk to you while they are forming an opinion. They want to talk to you after.

Which means that by the time a conversation happens, most of the opinion is already built. It was built out of whatever was findable, in whatever shape a machine could compress.

Where you are being read now

OtterlyAI, a company that tracks AI search, looked at 1.3 million citations of LinkedIn content across six AI assistants between January and June of this year. It is a vendor study, not LinkedIn’s own data, so hold the numbers loosely. The direction is still worth knowing.

In their sample, LinkedIn is now the third most cited social source in AI answers, behind Reddit and YouTube. Its share of those citations moved from 7.8 percent in January to 11.7 percent in May.

Two details inside that study matter more than the headline.

Content published by a named person was cited far more than content from company pages. Named people accounted for 91.7 percent of the citations they counted.

And long form was cited roughly 2.8 times more often per piece than short posts.

Put those together and the picture is uncomfortable for anyone whose entire public presence is short posts under a company banner.

The reader you will never meet

I run LinkedElla, a private positioning and ghostwriting practice for senior executives and founders, most of them in the United States. Almost everyone who comes to me arrives with the same problem. The authority is real, the public record of it is not.

Which is where a second version of all this begins, and that one is older than AI.

The Edelman and LinkedIn research on B2B thought leadership, built on responses from close to 2,000 decision makers, keeps finding that the people who shape a purchase are often not the people in the meeting. They call them hidden buyers. More than 40 percent of deals stall on disagreement inside the buying group, in a room you are not in.

Someone has to be able to explain you accurately when you are not there to correct them.

So there are now two intermediaries between your work and the decision. A machine that summarizes you, and a colleague who repeats you. Neither of them has your context, your tone or your caveats.

What survives being compressed

This changes what is worth writing, and it is not a question of reach.

Some things survive a summary and some things evaporate.

  • A claim survives. An atmosphere does not.
  • A number survives. An adjective does not.
  • A boundary survives. A benefit does not.
  • A named situation survives. A general principle rarely does.

Most executive content is written to sound right when read in full. Very little of it is written to stay true when cut to three lines by something that does not know you.

This is not an argument for saying more. It is an argument for saying fewer things more precisely, which is a different discipline. I have written before about why not everything worth saying belongs on LinkedIn.

The repeatable sentence

Here is the exercise. It takes an hour and it is worth more than a month of posting.

1. Write the claim you want repeated

One sentence. No metaphor, no wordplay, nothing that needs the paragraph around it. If a colleague repeated only this sentence in a meeting, would it still be accurate.

2. Attach something concrete to it

A number, a named condition, a specific situation. Not because numbers impress, but because they survive. “We work with founders after the first institutional round” survives. “We help ambitious leaders grow” does not.

3. Say who it is not for

Exclusions compress better than promises. They are also harder to fake, which is exactly why a reader trusts them and a summary keeps them.

4. Put it somewhere long

One properly argued piece, with your reasoning and your evidence in it, does more for how you are represented than eight short posts. Short posts carry it to people. The long piece is what gets quoted back to them.

5. Keep the sentence the same everywhere

Headline, About section, article, post, the way you introduce yourself on a call. Not identical wording, the same claim. Every variation gives the summary another version to choose from, and it will not choose the one you like best. I wrote about the headline itself in Your LinkedIn headline has two readers now.

Two tests, three weeks apart

First test. Ask someone who knows your work, but does not work with you, to say what you do in one sentence. Do not help them. What comes back is your real position, not the one on your profile.

Second test. Ask an AI assistant what you work on and who you work with. Read the answer as information about your public record rather than as a verdict. If it is vague, your record is vague. If it is wrong, something findable is telling it that.

Neither test is flattering. Both are cheap, and both tell you more than a month of engagement numbers.

The part worth keeping

Buyers have not become less interested in people. Gartner found that 69 percent still want a human to confirm what they learned, and that buyers who feel confident about a decision are twice as likely to call the deal a good one.

The conversation still decides it. It just happens later, and you no longer control what happened before it.

So the work is not to be everywhere. It is to be accurately repeatable by people and machines that will describe you without asking you first.

Questions people ask about this

Does this mean I should write long articles instead of short posts?

No. It means you need at least one long piece that holds your reasoning, and short posts that carry it to people. The posts create attention. The long piece is what gets quoted when you are not in the room.

How do I find out what an AI assistant says about me?

Ask it directly. What does this person work on, who do they work with, what are they known for. Ask two or three different assistants, because they read different sources. Treat the answer as a report on your public record, not as a judgment of your work.

What if the answer is wrong or empty?

Empty usually means there is nothing findable in your own words, so the fix is to publish one clear piece under your own name. Wrong usually means something findable is saying it, often an old bio, a stale company page or a title you no longer hold. Correct the source before you write anything new.

Do I need to post more often for this to work?

Frequency is not the lever here. Consistency of subject is. Eight posts on one claim will do more for how you are described than forty posts on eight subjects.

Who is this advice for?

Senior people with real authority and no public record of it. Founders, executives and specialists whose work is known inside a small circle and invisible outside it. This is what LinkedElla does, quietly and under agreement.