Key Takeaways
- Branded web mentions correlate with AI visibility about three times more strongly than backlinks (Ahrefs, 0.664 vs 0.218), so the link a digital PR campaign chases is now the weaker half of the placement.
- You cannot self-promote into an AI recommendation: Google cites a brand's own 'best of' listicle but leaves that brand out of the answer 69% of the time, so the mention has to come from a third party.
- Being cited is not the same as being named: 61.7% of AI citations never say the brand's name in the answer, so push placements to name you in the prose, in context, where buyers and models both read.
Digital PR earns the thing AI search runs on: your brand, named, in someone else's content
Most digital PR advice still ends at the link. Run a data study, pitch journalists, earn editorial coverage, count the referring domains. The coverage is real and the links are good. But the metric is dated. The backlink was always standing in for something the link itself could not measure: a credible third party deciding your brand was worth pointing to. AI search has made that underlying thing visible and measurable, and it is the brand mention, linked or not.
So the 2026 question for a B2B program is not "how many DR-60 links did this campaign earn." It is "how often is our brand getting named, in context, in the places buyers and AI engines read." That single shift changes which placements you chase, what you ask a journalist for, and how you report the work. The rest of this post walks the change and what to do about it.
The backlink was a proxy, and the proxy is now the weaker signal
Digital PR ran on domain rating and referring domains for a simple reason: links were the measurable artifact. You could not quantify "a respected publication vouched for us," so you counted the link it left behind. For classic Google ranking, that proxy held up well enough.
AI engines read differently. They assemble answers from prose, transcripts, and discussions, and they decide which brands to name based on how often and how widely those brands appear across that text, not on the link graph underneath it. Ahrefs studied 75,000 brands and found that branded web mentions correlate with AI Overview visibility at 0.664, while backlinks come in at 0.218. That is close to a three-to-one gap in favor of the mention. The same study found the top quartile of brands by web mentions averaged 169 AI Overview appearances against 14 for the quartile below them.
Read that as an operator and the conclusion is uncomfortable for link-first PR: the part of a placement you were chasing is the part that matters least to AI. A piece of coverage that names your brand in the body text but carries no dofollow link still does the AI-visibility work. A footer link with no brand name in the prose mostly does not. The link is now the bonus and the mention is the asset.
When the search does not send a click, being named is the only way to be present
The reason this matters more every quarter is that fewer searches end in a visit to your site at all. SparkToro and Similarweb found that in the first four months of 2026, 68% of US Google searches ended with zero clicks, with AI Overviews now appearing on more than 20% of searches and cutting click-through by close to 60% when they show.
If the answer resolves the query inside the results, the only form of presence left is being named inside that answer. A backlink cannot buy that seat, because the answer often has no outbound link to give and the click never happens. The brand that gets mentioned in the generated response is the one the buyer walks away remembering. Digital PR, done for mentions rather than links, is the supply line for those mentions. This is the off-site half of a working Demand Engine, and it is covered in more depth in our breakdown of how to rank in AI search.
You cannot self-promote your way in, so the mention has to come from someone else
For two years the cheapest workaround was to publish your own "best [category]" listicle and rank yourself first. It worked because the models had no good way to separate self-praise from genuine standing. That window is closing. Lily Ray's team at Amsive reviewed 100 B2B "best of" queries and found that when a brand published its own self-promotional listicle, Google's AI surfaces cited that listicle but left the brand out of the actual recommendation 69% of the time. Google cites your page, then recommends the established competitors named inside it. As the analysis put it, the recommendation is now anchored to how much of the web is talking about you, not how many times you have called yourself the best.
This is the structural case for digital PR specifically, over more owned content. The signal AI rewards is a third party saying your name, and you cannot manufacture that on your own domain. Analyses of AI citations consistently show the large majority come from external sources rather than the brand's own site. Owned content keeps you eligible to be cited; earned mentions are what get you recommended. A program built to produce other people's coverage is the only one that supplies the signal that actually moves the answer.
A citation is not the same as a brand mention
Here is the trap inside the good news. Being used as a source is not the same as being named. Semrush and Kevin Indig studied this gap and found that 61.7% of AI citations are "ghost citations": the page is used as a source, but the brand name never appears in the answer the reader sees. Across their data the citation rate ran nearly double the mention rate. The study also found comparative content ("best," "vs," "recommend") produced about 2.4 times more brand mentions than informational content.
For a digital PR program the instruction is concrete. A placement that links to your study but describes it as "one analysis" is half a win. A placement that names your firm as the source of the finding, in the sentence, is the full one. So brief your spokesperson and your data so the natural way to reference them is by name, pitch into the comparative and recommendation contexts where naming happens, and treat an unlinked mention that says your name as more valuable than a linked one that does not. The named mention in context is the unit of work.
The practical version of this is naming your dataset and your point of view so a writer cannot reference them generically. A study called "the 2026 B2B Pipeline Benchmark from [your firm]" gets attributed by name; "a recent survey" does not. The same logic applies to a proprietary framework or a number only you publish. When the finding carries your name as part of its identity, the mention comes along with it whether or not the journalist adds a link.
The surface that counts is wider than journalist placements
Classic digital PR aimed at journalists and column inches. The mention signal lives in more places than that. Ahrefs expanded its study across ChatGPT alongside Google's AI Mode and AI Overviews, and found YouTube mentions were the single strongest correlate of AI visibility at roughly 0.737, ahead of branded web mentions, with backlinks again near the bottom and raw content volume showing almost no relationship at all.
For a B2B company the practical map is to go where buyers and models both read: industry publications and podcasts, expert roundups, review platforms like G2 and Capterra, relevant Reddit and Quora threads, and video where your category gets discussed and transcribed. The job here has little to do with broad consumer awareness. It is being the named expert and the cited source inside the specific corner of the web your buyers research in. That is the same job as earning a citation in an answer engine, which is why digital PR and answer-engine work now run as one motion rather than two budgets. If you want the on-site complement to this off-site work, our SEO and AEO checklist covers the page-level moves that keep you eligible to be cited in the first place.
What to measure instead of domain rating
Replace the link scoreboard with a mention scoreboard. Track how often your brand is named across the surfaces above and whether each mention sits in a context your buyers actually care about. Track your share of those mentions against the competitors who keep showing up in the same answers. And separate "named" from "merely cited," because the ghost-citation data says those are different outcomes that a single citation count will blur together. For a B2B operator the work earns its keep by making you the brand the answer names when a buyer asks who does this, which puts you on the shortlist before a form ever gets filled. That ties digital PR to the one number the rest of the system reports on. Booked sales conversations is that number, which is why we treat it as part of the Demand Engine inside one connected growth system rather than a standalone PR line item.
I ran marketing for a 4x Inc. 5000 firm from startup to exit, and the pattern held long before AI search named it: the coverage that compounded was never the link we negotiated, it was the moment a credible outsider described what we did in their own words. AI engines have turned that instinct into a measurable signal. Run digital PR for the named mention, wire it to the same number as your search and outbound work, and you are building the off-site layer of a single growth system instead of buying links and hoping they move something.
