Does Digital PR Work for AI Search Visibility?

Does digital PR actually work for AI search visibility?

As of September 2026, yes, but not for the reason most agencies sell it. Digital PR works because it puts your brand inside third-party pages that AI engines already trust and quote. The link is secondary. Being described accurately, on a source that gets cited, is the real product.

We keep getting asked to plug a PR retainer into a website project, and the pitch is almost always about domain authority. That framing is a decade old. The question worth asking in 2026 is narrower: will this placement put a correct sentence about your company somewhere a language model will read?

Here is how we think about it, with the data that is actually published rather than the data agencies imply.

What is digital PR, and what is it not?

Digital PR is earning coverage on other people's websites through something newsworthy: original data, a strong point of view, an expert comment, a launch that matters. The output is an article on a publication you do not control.

It is not link building with a press-release coat on. It is also not guest posting at volume. Google's spam policies draw that line explicitly, and we will get to the exact wording below.

The useful distinction for AI visibility is between a placement that describes your company and a placement that merely links to it. A single sentence that correctly says what you do, on a page that gets quoted, does more for an AI answer than ten no-context backlinks.

Which sources do AI answer engines actually cite?

Semrush published a study on 10 November 2025 analysing 230,000 prompts across ChatGPT Search, Google AI Mode and Perplexity, taking weekly snapshots between 14 July 2025 and 12 October 2025 across thirteen weeks, and tracking over 100 million AI citations. The shape of the answer is that citation sources are unstable and platform-specific.

On ChatGPT, Semrush recorded Reddit citations falling from close to 60 percent of prompt responses in early August to around 10 percent by mid-September, and Wikipedia falling from roughly 55 percent to under 20 percent in the same window. On Google AI Mode, LinkedIn appeared in nearly 15 percent of responses while Wikipedia appeared in around 2 percent. Perplexity was steadier, with Reddit, LinkedIn, NIH, Microsoft and Google leading throughout.

Read that as a warning about strategy. Any plan built on one platform's favourite source can be undone in six weeks by a change you never see announced. The durable version is to be present across many credible sources rather than to chase whichever one is currently dominant.

Why do mentions matter more than links here?

Because a language model reading a page does not need an anchor tag to learn that you exist. It needs a sentence. An article that says a company does a specific thing for a specific buyer gives a model a fact it can repeat. A bare hyperlink on the word "here" gives it nothing.

This is the biggest practical shift from classic link building. In classic SEO, the link carried the value and the sentence around it was packaging. In AI answer engines, the sentence carries the value and the link is packaging. Our piece on unlinked brand mentions goes deeper on why that inverts a lot of agency pricing.

It also changes what you should ask a journalist for. Asking for a followed link is a negotiation. Asking them to describe your product accurately is usually just accuracy, and most writers will take the correction.

Does Google's link spam policy apply to digital PR?

It applies squarely, and it is worth reading the actual text. Google defines link spam as "the practice of creating links to or from a site primarily for the purpose of manipulating search rankings." The word doing the work, again, is primarily.

Google's examples name the tactics that digital PR agencies sometimes drift into: "Buying or selling links for ranking purposes," including exchanging money, goods or services for links; and "Advertorials or native advertising where payment is received for articles that include links that pass ranking credit, or links with optimized anchor text in articles, guest posts, or press releases distributed on other sites."

There is a clean way through. Google states that paid placements are fine when qualified: "It's not a violation of our policies to have such links as long as they are qualified with a rel='nofollow' or rel='sponsored' attribute value to the a tag." So sponsorship is allowed. Sponsorship dressed up as earned editorial is not.

Are press releases worth it for AI visibility?

A press release is a distribution mechanism, not coverage. Google's policy explicitly flags links with optimised anchor text in press releases distributed on other sites as link spam. That kills the old wire-service link play outright.

What a release still does is create a canonical, quotable statement of a fact on your own newsroom. If you launched something, a clear dated page on your site saying what launched is genuinely useful to a model trying to answer a question about you. The wire distribution is the part with limited value.

So publish the release on your own domain, keep it factual, date it, and let journalists pick it up on merit. Our view on the wider question sits in whether press releases still help SEO.

How is this different from traditional link building?

The target list changes. Traditional link building chased domain metrics. AI visibility work chases sources that the engines demonstrably quote for your category, which for B2B software often means review platforms, industry publications, community threads and analyst-style write-ups rather than general news.

The asset changes too. Link building could run on outreach alone. Earning a citation usually needs something quotable: a number you generated, a method you can explain, a position someone disagrees with. Without that, there is nothing for a writer to lift.

And the timeline changes. A link's value used to compound slowly and quietly. A mention on a frequently-crawled page can show up in AI answers much faster, and can also disappear faster when the page changes. It is a less stable asset than a backlink ever was.

What does a realistic programme look like for B2B?

Start with accuracy, not reach. Find every existing page that already describes your company and check whether the description is right. Wrong descriptions on well-trusted pages are actively harmful, because they are exactly what a model will repeat back to a buyer.

Then build one genuinely original asset a quarter. Original data beats commentary for pickup, because a writer can cite a number but cannot cite an opinion without giving you a whole paragraph. If you have a dataset nobody else has, even a small one, that is the lever.

Then place expert commentary where your buyers already read, and make sure the bio sentence is precise. In our work the bio line is the most under-optimised asset in the whole programme. It is often the exact string a model will reuse.

How do you measure whether it worked?

Not by referral traffic, and this is where expectations need resetting. Pew Research Center analysed 68,879 unique Google searches from 900 US adults tracked through March 2025, with results collected between 7 and 17 April 2025. It found users clicked a traditional search result on 8 percent of visits where an AI summary appeared, against 15 percent where none did.

The citation click numbers are starker. Pew found users clicked a link inside the AI summary itself on just 1 percent of visits to pages carrying one, and that 26 percent of pages with an AI summary ended the browsing session, against 16 percent for pages with only traditional results. Visibility inside the answer is not a traffic channel in any conventional sense.

So measure share of voice in the answers themselves. Run your real buyer questions through ChatGPT, Perplexity, Claude and Google AI Overviews on a schedule, record whether you appear and whether the description is right, and track that over months. Our guide on getting cited by AI search engines covers the mechanics.

Where does this go next?

Our honest read is that digital PR is being rediscovered by SEO teams for the right reason and sold for the wrong one. The value is accurate presence across many trusted third-party pages. The pitch is still usually authority scores, because that is easier to put on a slide.

The teams that will do well are the ones that treat this as reputation work with a measurement loop attached, and that accept the Pew reality: winning the answer rarely wins the click. It wins the shortlist, which is the thing that actually matters for a considered B2B purchase.

If you want a second opinion on whether a PR programme is worth it for your situation, we are happy to walk through it with you. Find us at phoenix.studio.