AI Search Cites People, Not Your Brand Page (2026)

This morning I pulled the live Google results for three queries our category actually gets asked, and read who was being cited. On "b2b creator marketplace", the first six organic results were LinkedIn's own newsroom, a Digiday article, a LinkedIn post from a marketer, Business Insider, and two agency blogs. Google's AI Overview cited eleven domains. On "best b2b influencer marketing platform", the second organic result was a Reddit thread, and the AI Overview drew on six sources, one of which was that same Reddit thread.

I run a company in this category. We have a comparison page, a pricing page, a blog with fifty articles. On the query that describes literally what we sell, the answer being assembled for the buyer was built mostly out of journalists, a LinkedIn post by someone who does not work here, and a Reddit thread.

I do not think this is a temporary artefact of a young feature. I think it is the structure of the thing, and that it quietly changes what a B2B distribution budget is for.

The pages you control are the ones an answer engine has the least reason to trust

The aggregate numbers say the same thing my three queries did. Citation-index research covering hundreds of millions of citations across ChatGPT, Google AI Overviews, Perplexity, Gemini and Claude consistently puts Reddit, YouTube and LinkedIn at the top of the most-cited domains, with Reddit accounting for something like 46.7% of social citations on Perplexity and 21% on Google AI Overviews. A small group of community, video and encyclopedia sources absorbs a large share of every citation these systems produce.

The obvious reading is that these sources are big and well-crawled. I think the more useful reading is about what an answer is trying to be.

When a model answers "which platform should I use", it is not writing a directory entry. It is performing the job a trusted colleague performs: giving you a shortlist with reasons. A shortlist assembled entirely from the vendors' own marketing pages is worthless, and the systems behave as if they know it. So the vendor's page gets used for the facts about the vendor — what it costs, what it does — and something else gets used for the judgment about whether the vendor is any good.

That "something else" is almost always a place where identifiable humans said things. This is the part that matters: your owned pages are structurally disqualified from the most valuable slot in the answer, and no amount of improving them changes that. They can be excellent and still lose, because the thing they are losing on is not quality. It is independence.

I want to be careful not to overclaim. Citation-share studies measure which domains get cited in aggregate. They do not show that a LinkedIn post about your product causes your product to be recommended, and nobody's tooling can attribute that today. What they establish is where the citation supply lives. That is enough to reason about, and not enough to promise anything.

The wrong conclusion is to go manufacture the independence

The moment a growth team sees "Reddit is 40% of citations", the plan writes itself: get us into the threads. I have watched a few versions of this and I think it is a mistake, for a reason that is more interesting than "you'll get caught".

You will get caught — moderators are unusually good at this, and the platforms have every incentive to protect the signal. But the deeper problem is that the tactic destroys the asset it is trying to exploit. Community sources are cited because they read as disinterested. A brand that manufactures disinterested-sounding endorsement is producing a counterfeit of exactly the property that made the source valuable. If it worked at scale, it would stop working, because the sources would stop being worth citing.

The same logic applies to the softer versions: the ghostwritten "practitioner" post, the employee advocacy programme where twelve people publish the same paragraph on the same morning, the review left in exchange for a gift card. These are not cheating so much as they are self-defeating. They produce volume in the citation layer and degrade the credibility of the channel that the volume was supposed to buy.

What does not self-destruct is the boring version: a real practitioner, with their own audience, who has actually used the product, saying what they think under their own name, disclosed as sponsored. It survives scrutiny because there is nothing to discover. The disclosure is not a compliance cost — it is the thing that makes the endorsement legible as an endorsement rather than as an attempted deception.

This is what creator-led distribution already is, bought for the wrong reason

Here is where I will make the argument that this article exists for.

Most B2B teams that buy creator posts buy them as a reach channel. They evaluate creators on follower count, they measure the campaign on clicks and attributed pipeline, and when the clicks are mediocre they conclude the channel is mediocre. I understand the instinct; we sell into it, and our own marketplace is full of buyers who think in exactly these terms.

I think the scoreboard is now wrong. A sponsored post from a credible practitioner produces three things: the direct audience response, a durable public artefact under a real name, and a piece of the citation supply for the questions your buyers will ask a model six months from now. Teams measure the first, occasionally the second, and almost never the third — even though the third is the one that compounds and the one that a brand page cannot buy for itself at any price.

Two consequences follow, and they cut against how the category is usually sold.

The first is that the credibility of the creator matters more than their size, which is convenient for me to say and also what our own data suggests. Follower count explains only about 29% of the variance in what B2B creators on our marketplace list as their per-post price [Naano Index snapshot, n=769 creators, 11 August 2026 at 19:05 UTC] — the market is already pricing something other than reach. If you are buying citation supply rather than impressions, a 4,000-follower security engineer whom other security engineers quote is worth more than a 60,000-follower generalist, and the price difference usually runs the other way.

The second is that one post is close to worthless and a habit is not. A single sponsored mention is a single artefact. What gets absorbed into the way a category is described is repetition across independent voices over time. This is genuinely expensive and genuinely slow, and I would rather say so than pretend a campaign fixes it.

What I would actually tell a B2B marketing lead to do

Keep the owned pages. They still do a job — a real one. Once a model has surfaced your name because a practitioner mentioned you, the follow-up questions are factual: what does it cost, who is it for, how does it compare to the alternative. Those answers should exist on your site, stated plainly, with numbers you can stand behind. Publishing real prices is the single highest-leverage thing most B2B companies could do here and almost none of them do it, which is why the AI answers in our category are so vague about cost.

But stop expecting those pages to win the recommendation. Budget separately for the layer you cannot own: practitioners who will talk about you in public, under their own names, more than once, with disclosure. Choose them for whether their peers take them seriously rather than for audience size. Accept that you will not be able to attribute it cleanly, and decide in advance that you are buying presence in the citation supply rather than a trackable click — because if you measure it on clicks you will kill it in one quarter, and the clicks were never the point.

I am aware this is a convenient thesis for a company that sells access to B2B creators. Treat it accordingly, and go check the SERPs in your own category before you believe me. It takes ten minutes, and whatever you find will be more persuasive than this article.

If you want the market view first, our comparison of B2B creator platforms and what each is actually for is the honest version, and what creator-led growth means in B2B covers the fundamentals. When you are ready to buy the practitioners rather than the reach, tell us what you are launching and we will find them.