PR’s Role in AI Search

One of the more interesting shifts we've watched over the last two years is how quickly consumers have become comfortable asking AI systems questions they would previously have typed into a search engine.

Sometimes those questions are simple:

• What's the best standing desk?
• Which smartwatch should I buy?
• Who makes the most reliable electric bike?

Sometimes they're much more nuanced and come from longer prompts. Traditional search gave users a list of links and asked them to do the work. AI systems increasingly provide an answer. They summarize and recommend.

That subtle shift has significant implications for brands.

When a consumer receives ten search results, there is still an opportunity to compete for attention. When an AI system produces a single recommendation or a shortlist of suggested companies, the competitive landscape becomes much narrower. Visibility is no longer about being present somewhere on the page.

Many organizations are still thinking about discoverability in terms of rankings, traffic, and content volume. The more important question is rapidly becoming authority. Why does an AI system trust one source over another? Why does it mention one company and ignore five others operating in the same category?

Those questions sit at the heart of GEO. And the numbers show they lead directly to PR.

For decades, PR has operated on a fairly simple premise: what other people say about your brand carries more weight than what you say about yourself.

Nielsen has found, for decades, that consumers trust independent reviews more than advertisements. Journalists trust evidence more than marketing claims. Investors trust third-party validation more than company messaging.

AI systems appear to be following the same logic.

The same researchers at Princeton, Georgia Tech, and IIT Delhi found that AI engines consistently favour authoritative third-party sources over brand-owned content. Since that research was published, a growing body of evidence has pointed in the same direction.

Media coverage creates a network of authority signals that helps AI systems understand which brands deserve attention. Every article, review, interview, and expert quote contributes another data point. Over time, those data points begin forming a pattern.

LLMs are basically powerful pattern matchers.

Any individual statistic can be challenged.

What becomes difficult to ignore is when multiple studies begin arriving at the same conclusion from different directions.

That is increasingly what we are seeing around AI search visibility.

According to research highlighted in our report, 82% of AI-cited links come from earned media sources. Across major AI platforms, 94% of citations come from non-paid sources. Brands are 6.5 times more likely to be cited through third-party sources than through their own domains.

The findings become even more interesting when you look at journalistic content specifically.

Recent research from Muck Rack suggests that 27% of all AI citations are journalistic. For time-sensitive queries, that figure approaches 50%. In practical terms, that means reporters, publications, and editorial coverage are playing an outsized role in shaping how AI systems answer questions.

One of the strongest pieces of evidence comes from research conducted by Stacker and Scrunch. Brands saw a 325% increase in AI citations when content was distributed through third-party news outlets rather than published solely on their own websites.

That finding should make every communications leader pay attention.

For years, earned media has been valued because it influences audiences. Today, it appears to influence another audience as well: the systems increasingly responsible for helping those audiences make decisions.

One of the least appreciated aspects of AI visibility is how quickly it decays.

Many brands still approach earned media as a campaign activity. They secure coverage, celebrate the results, add the logos to a slide deck, and move on to the next initiative.

That approach made more sense when the primary audience was human.

AI search instead rewards consistency.

Research cited in our report found that the most common time for content to be cited by AI systems is shortly after publication. Roughly half of AI citations reference content published within the previous eleven months. Other studies suggest that the useful lifespan of many citations is measured in weeks, not years.

This presents a challenge for brands that treat visibility as a one-time achievement.

A company may have secured outstanding coverage three years ago. It may have won awards, launched successful products, and built a strong reputation within its industry. None of that guarantees continued visibility if newer signals are constantly entering the information ecosystem.

Authority requires ongoing maintenance.

We've spent years explaining to clients that media coverage does more than generate traffic.

A strong article can influence how customers perceive a company. It can affect investor confidence. It can strengthen retailer relationships. It can create social proof that improves conversion rates long after publication.

AI search has expanded the audience for that influence.

The same article that helps a consumer evaluate a product today may help an AI system evaluate that same product six months from now. The executive interview that establishes thought leadership within an industry may become part of the information pool AI systems draw from when explaining that industry to users.

This is one reason we see GEO and PR as deeply interconnected disciplines.

Many of the signals AI systems appear to value most are direct outputs of effective PR:

• Media coverage.
• Product reviews.
• Executive thought leadership.
• Expert commentary.
• Industry research.
• Conference speaking.
• Podcast appearances.
• Category-defining narratives.
• Third-party validation.

None of these activities were originally designed for AI search. What has changed is that AI systems are now consuming many of the same signals humans have relied on for years when evaluating credibility.

This is particularly powerful during periods of category creation.

When a new technology category emerges, somebody has to explain it. The companies that teach AI systems how to frame a category get outsized influence and visibility.

We've written previously about Share of Explanation, the idea that influence increasingly belongs to the brands that shape understanding rather than simply appearing in answers. AI search amplifies this phenomenon. Visibility matters. The ability to shape how a category is understood may matter even more.

Keychron was not the first mechanical keyboard company. The category existed long before the brand entered the market. What Keychron did exceptionally well was establish itself as one of the defining voices within that category through sustained investment in earned media, reviews, community engagement, and category education.

As the category grew, so did the volume of third-party content surrounding the brand. Journalists covered it. Reviewers tested it. Influencers recommended it. Consumers discussed it. Over time, Keychron became strongly associated with the category itself.

That association now extends into AI search.

When consumers ask AI systems about mechanical keyboards, Keychron overwhelmingly appears in the answers in the top position. While there are many factors influencing those recommendations, the company's earned media footprint is difficult to ignore.

We've seen similar patterns emerge across categories such as e-bikes, drones, fitness technology, and smart home devices.

Technology companies should view these developments as both a challenge and an opportunity.

The challenge is that visibility can no longer be treated as a purely owned-media exercise. Building authority requires participation in a broader ecosystem of journalists, reviewers, analysts, creators, researchers, and industry experts.

The opportunity is that many competitors still have not adjusted.

A significant number of brands continue approaching AI visibility as though it is primarily a publishing problem. They focus on producing more content without asking a more important question: who else is talking about us?

That question matters whether you are entering the U.S. market, launching a new category, defending share in a crowded industry, or attempting to establish yourself as a category leader.

If you are curious about a diagnostic, reach out to info@properpropaganda.net and we can chat about getting you a snapshot.