PR and SEO: Why Your PR Isn't Showing Up in Search

Press coverage moves SEO rankings and AI visibility when it’s built to answer the questions buyers ask. As of 2026, most campaigns aren’t, which is why a company can spend $15,000 on placements and see zero movement three months later. Google shows the same results, but ChatGPT recommends a competitor.

The disconnect isn’t effort or reach. It’s that traditional PR is built to produce coverage, and search systems, both Google and AI models, need something more specific. A placement earns a ranking or a citation only when it contains the kind of detail those systems can attach to a query, and most coverage doesn’t.

PR Is the Umbrella Over Search and AI Visibility

Most companies budget for public relations, search optimization and, more recently, answer engine optimization as three separate lines. Treating them as separate entities is the first mistake.

Edgar Li, founder of Presscart, a media placement company that puts client stories in front of a contributor network of more than 1,700 publishers, treats the three as one discipline. “PR is more of a methodology, an approach to creating stories,” he says. Positioning, story creation and the decision about where that story lives all sit under it. Search performance and AI citations are outcomes of those decisions, not parallel disciplines with budgets of their own.

That framing matters most for AEO, which Li splits in two. “A big part of an AEO strategy requires both on-site content and off-site content,” he says. On-site content gives a model something to read about the company. Off-site coverage gives it corroboration from somewhere other than the company itself.

The mechanism behind the off-site half is old. Outside links function as a vote of confidence: when other sites cover a company and link to its pages, Google reads that as evidence the domain carries weight. Li compares it to a referral. People take the meeting with the person their friends vouch for.

What changed is who else counts the votes. An Ahrefs study of 75,000 brands found that branded mentions across the web correlated with visibility in Google’s AI Overviews at 0.664, against 0.218 for backlink volume, a roughly threefold gap, with the three strongest signals all coming from off-site sources.

The direction still tracks what Li describes: “With PR now, you’re not just doing PR for other people: you’re also doing PR for AI models.”

Why PR and SEO Campaigns Miss the Citation

Coverage that fails to produce citations usually fails for one of two reasons, and the two aren’t the same problem.

• No goal was set. The content went out without a keyword to target or a claim a model could repeat, so nothing about it was ever built to earn a citation.
• Earned press worked as designed. An independent outlet writes to its own angle, which is what makes the coverage credible and also what leaves the company’s positioning out of it.

The first shows up as enthusiasm without a plan. “A lot of the time, people get excited, but they’re not very thoughtful or intentional about the process,” Li says. “They’ll jump into it without any understanding of how these things work, or why things get cited.” The sequence is easy to recognize. A company generates an article in a chat with ChatGPT or Claude, publishes it somewhere, and waits for citations that never come.

“If you’re not setting a clear goal for your content, you’re most likely not going to reach that goal,” Li says. Content built that way can run in a dozen outlets and still leave a search engine with nothing specific to attach to the company, which is why press spend and rankings so often move on separate tracks. Presscart builds each placement backward from the citation it’s meant to earn, which is the step most campaigns skip.

The second is no mistake at all. An objective third party writes the story to its own guidelines and its own perspective, so how the company wants to be described never enters the calculation. “Their first priority is getting the right story out,” Li says.

That priority is why the coverage so often does nothing for search. Companies, in Li’s words, “miss the opportunity to get cited.” The article exists, and it may even be good journalism, but it just wasn’t built to answer the question a buyer types into a model, and it describes the company on terms the company never picked. The counterweight is content a brand controls, written from the start with citation as the goal.

How to Judge an Outlet, and Whether It Worked

Buyers tend to evaluate outlets by a single number. Li works through three, in order.

• Authority. Domain rating is the working proxy for how much credibility a publication carries, and it is the filter most people stop at.
• Relevance. The outlet should cover the industry the brand actually competes in, which is what determines whether the citation lands anywhere useful.
• Audience. The placement has to sit somewhere the buyer already goes, since a story nobody in the market reads earns nothing.

The first filter is the one people over-trust. “From a technical perspective, you want to go toward authoritative publications, and a proxy for that is the domain rating, or domain authority,” Li says. He treats the number as a general benchmark and not a verdict, since AI models do cite lower-authority sites.

The second filter is what a high score alone can’t buy. “If you’re a chemical industrial firm, it doesn’t make sense for you to be cited in a fitness magazine,” Li says. With the exception of general news outlets, broad enough that most companies can sensibly appear in them, a strong publication covering the wrong subject returns a link and little more.

Not all placements carry the same weight, either. Feature articles and guest bylines in niche publications earn both backlinks and AI citations because they contain substantive claims and detailed positioning. Roundups and directory listings build backlinks but rarely citations, listing a company as one option without explaining what it does or who it serves. Press releases work only when distributed to outlets that republish with attribution and context, not syndicated to link farms.

This is the test Presscart applies to its own network of more than 1,700 publishers, which it grades for quality against strict editorial standards used by outlets including VentureBeat and USA Today, then filters down to the publications whose subject matter overlaps with the buyer intent keywords a given client is trying to win.

The third filter is where the other two collide. What a model treats as a legitimate source is roughly what a person would, which puts audience and authority in the same column instead of opposite ones. “These models, to a certain degree, are trying to replicate how a human searcher would work,” Li says.

Measurement follows the same logic. “Previously, in the pre-AI world, just with SEO, you could measure your positioning through backlinks,” Li says. “But now, with AI taking over search ranking, the system has become more complex.”

Take a restaurant going after “best hamburgers in New York City,” a phrase every burger place in the city wants. Winning it takes two things at once: a page on the restaurant’s own site built around that phrase, and coverage that uses the phrase and links to that page, which is what digital PR services like Presscart arrange.

The two halves tell Google different things. The restaurant’s own page establishes what the site is about. The incoming link establishes that somebody credible thinks so too, on that same subject. Google needs both readings to place the restaurant in a competitive ranking, which is why either half alone stalls. A page nobody links to has nothing vouching for it, and a link from a story that never mentions burgers vouches for nothing in particular.

AI search reads the same two signals and adds a third. A model doesn’t just count who links to whom; it reads what those pages say and decides whether the picture holds together.

“It’s got a variety of different people from different outlets. It’s got people mentioning it on Instagram or Reddit,” Li says. “And from there, the AI is going to reason which option is the best one to recommend.” Companies track whether the brand appears in AI responses to target queries, whether new backlinks appear from relevant publications, and whether rankings shift 60 to 90 days post-placement.

Consistency across independent sources is what tips that judgment, which is why a scattered footprint underperforms a coherent one.

Visibility Is Half the Job

Getting surfaced and getting described correctly are two different objectives, and companies routinely solve for the first while ignoring the second.

• Visibility. Whether a model surfaces the company at all when someone asks a question in its category.
• Positioning. What the model then says the company is, which is the half that gets skipped.

“First, a brand needs to be visible to AI and to other humans,” Li says. “And then second, it has to make sure search engines and AI are understanding the right things about it. That’s the second part that gets ignored. Okay, now you’re being seen, you’re being heard. But what is actually the message you’re delivering?”

Correcting this starts on property the company controls. Models spend a lot of time going onto your website to understand the content a company’s pushing out, working out what the source is and who it serves before answering a question about it. A site that says this clearly gives every off-site mention something consistent to corroborate.

Li also points out how brand-owned newsrooms are the emerging form of that work, and Presscart is now applying its publisher grading process to them, on the theory that a company’s own domain should meet the same editorial bar as the outlets it pitches.

The through-line is continuity. Companies still have to decide what they want to be known for, get credible outlets to say media coverage that counts, and make sure their own pages say the same thing. Presscart’s approach to brand-owned newsrooms applies that sequence to content published on a company’s own domain.

From Coverage to Citations

The mechanics are new only in who reads them: a page that states plainly what a company does, coverage from outlets whose readers overlap with the buyer, and a description consistent enough across sources that a model arrives where a person would. What fails is seldom the spend, and almost always the missing decision about what the coverage was supposed to say.

Presscart’s position is that this is a new way of doing PR running on principles that haven’t moved, which leaves the old question in place: what should the company be known for, and does everything published about it say the same thing.