In AI search, branded mentions beat backlinks 3-to-1
A B2B buyer opens ChatGPT, types "who are the leading vendors for X," reads one synthesized answer with three or four named sources, and walks away with a shortlist. No demo form. No Google search. No call to your SDR. By the time anyone in your CRM hears the buyer's name, the consideration set is already closed.
This is not a fringe behavior. Gartner found that 67% of B2B buyers now prefer a rep-free buying experience, and self-serve research increasingly starts with an AI assistant rather than a search box. The first move in the buying journey stopped being a query and became a question to a model. Which means the question that actually decides your pipeline is no longer "do we rank on Google." It is "does the model name us, and whose words does it quote when it does."
The model has a strong, measurable preference for content with a human and an opinion behind it. The safe, anonymous, committee-approved register that most B2B brands default to is exactly the register it throws away.
Named humans get cited. "The Acme Team" gets dropped.
One of the more useful findings in current AI-citation research is about bylines. Content with a named, attributed author gets cited materially more often than the same content published anonymously. One analysis of ChatGPT and Perplexity put the gap at around 1.9 times more citations for named authorship. The point is not the exact multiplier, which varies by study. It is the direction, which is consistent: a name and a face beat "The Acme Team" every time.
The reason is not sentiment. A named author can be cross-referenced against other publications, podcasts, and social profiles, so the model can build a fuller picture of who is making the claim and whether other surfaces corroborate it (a pattern several AI-search analyses describe). Anonymous corporate content has none of that connective tissue. It floats, unverifiable, and the model treats it accordingly.
So the most common B2B publishing habit, running the blog under the company account because "it's not about any one person," is now a direct tax on your AI visibility. You are stripping out the exact signal the discovery layer rewards.
"Cold email is dead" is not citable. The number is.
Bylines get you considered. Specificity gets you quoted.
Models cite content that makes a defensible claim. "Cold outbound email open rates declined 34% in Q4 2025" is citable, because it names a number, a period, and an implicit method. "Cold email is dead" is not, because there is nothing to attach a citation to. The posts that move citation share name numbers, dates, methodologies, and explicit arguments. Most B2B content does the opposite, because it was written to make no one in the room uncomfortable. Language models penalize that register.
The instinct to hedge, to broaden, to keep every sentence safe enough that legal and three VPs would sign off, used to be the low-risk play. In a retrieval world it is the high-risk play, because hedged content gives the model nothing to grab.
It compounds with how retrieval actually works. When an engine answers from the live web, it builds a pool of candidate pages and re-ranks them on authority and how cleanly a passage answers the exact question. A self-contained sentence that states a specific claim is extractable. A paragraph of hedged filler is not. And on Google AI Mode, nearly every answer is a competition for about five citation slots, so "extractable" is the whole game.
Brand mentions, not backlinks, are the currency now
There is a structural reason a consistent POV beats a one-off campaign. Across 75,000 brands, Ahrefs found that branded web mentions correlate far more strongly with AI Overview visibility than backlinks do: 0.664 versus 0.218, roughly three to one. The brands earning the most web mentions pulled in up to 10 times more AI Overview mentions than the next quartile down.
Mentions accrue when a specific person keeps saying a specific thing across enough surfaces that the model starts treating it as a consensus position rather than marketing copy. That is just the description of someone publishing a real point of view, consistently, under their own name. The thing that used to be a soft branding nicety is now a mechanical input to the discovery layer.
Three unglamorous moves that change your odds
"Have a POV" stopped being advice about credibility. It became the closest thing B2B has to SEO for the AI era. The model is now the first reader, and it reads for a named human, a defensible claim, and a consistent track record across the web. The practical move is unglamorous: move bylines from the company account to actual people, make claims specific enough to be quoted, and publish the same point of view often enough that it shows up in more than one place. Your buyers are asking a machine who to trust. Right now the machine is reading everyone's content and deciding whose words are worth repeating. Generic, anonymous, hedged content does not just underperform anymore. It is invisible at the exact moment the shortlist gets built.