Only 8 percent of visits to Google search results pages featuring an AI summary resulted in a click on a traditional search result, compared with 15 percent of visits to pages without one. Pew Research Center reported the findings in July 2025, drawing on the browsing data of 900 US adults who shared their activity in March 2025 and an analysis of 68,879 Google searches.
The difference highlights a change in how people encounter information online. Search engines are no longer just directing users toward pages. They are increasingly presenting synthesized answers before users decide whether to explore the sources themselves.
For businesses, that raises a question about the purpose of public relations. If potential customers form an initial impression without visiting a company’s website, how should organizations communicate their news, and what can they reasonably expect publication to accomplish?
Baden Bower, a public relations agency led by CEO AJ Ignacio, works with businesses seeking visibility across established publications and digital channels. Its guide to getting press coverage discusses approaches to media exposure, including direct pitching, journalist-source platforms, and agency-assisted publication. These approaches sit within a wider industry debate about what publication can deliver when search engines and AI tools shape how audiences find information. An article can establish a public record, but whether readers find it, whether a search engine surfaces it, and whether an AI system cites it are separate questions.
AI summaries are changing the route from discovery to a decision
Pew’s research found that around 18 percent of Google searches in its dataset displayed an AI summary. The summaries frequently cited several sources, giving users a synthesized response rather than requiring them to compare every result independently.
The study also found that Wikipedia, YouTube, and Reddit were the most frequently cited sources, together accounting for about 15 percent of the sources listed in AI summaries. Established news outlets were only one part of a wider information environment, and only 1 percent of visits led to a click on a link inside the summary itself.
These findings challenge the assumption that appearing in a recognized publication automatically translates into online visibility. A story may be available on a publisher’s website without appearing in a particular search result or AI-generated response. The research does not establish that paid placements, press releases, or any other single content format are more likely to be cited.
For communications teams, the practical implication is to separate what they can control from what they cannot. They can prepare accurate announcements, provide relevant context, and keep company information consistent across public channels. They cannot dictate how search systems rank, summarize, or select that information.
Company news needs substance beyond the headline
AI-driven discovery is also changing how business buyers research suppliers. In its March 2026 survey of 1,076 business-to-business software buyers and decision-makers, G2 found that 51 percent began their software research with an AI chatbot more often than with Google. The result, published in The Answer Economy, suggests that conversational tools have become an important starting point for many respondents.
The survey also found that 85 percent said they thought more highly of a software vendor when an AI chatbot mentioned it in an answer. That figure reflects respondents’ reported perceptions. It does not prove that an AI mention independently creates trust or causes a purchase.
Still, it illustrates why businesses need public information that can withstand comparison and scrutiny. A potential customer researching a company may encounter its website, product documentation, customer reviews, press releases, and articles published by third parties. Each source offers a different perspective, and readers benefit when those distinctions stay clear.
A company-sponsored article should not be represented as independent reporting, and the presence of a company’s name in a publication does not by itself validate its claims. Baden Bower’s approach emphasizes predictability in the publication process, including identifying publication opportunities and setting expectations before a campaign begins. “We eliminated the uncertainty that has defined PR for decades,” Ignacio has said. Yet securing an article is only one stage of a communications effort, and the harder question, as the search data illustrates, is what happens after publication and whether the intended audience encounters the information.
Credibility depends on transparency as well as visibility
It is worth asking whether a paid placement offers the same kind of evidence as independent journalism. The answer depends on the format, how the content was produced, and whether the commercial relationship is disclosed. An article arranged by a company can carry useful information, but it does not hold the same evidentiary weight as reporting a journalist initiated and developed independently.
That distinction matters because visibility and credibility are not interchangeable. A company may secure a publication without reaching its intended audience. An article may draw attention without generating qualified inquiries. Search systems may exclude a page from an AI-generated answer even when the page is publicly accessible.
Companies should also be cautious about treating domain metrics or publication names as universal measures of value. A publication’s reputation may matter, but it cannot replace an assessment of audience fit, article quality, factual accuracy, and transparency. Nor does a high domain score guarantee search rankings, traffic, AI citations, or sales.
These limits do not make public relations irrelevant. They make the objective more specific. A communications effort should start with the audience a business needs to reach and the information that audience needs, rather than assuming publication alone will produce a commercial result.
As AI-generated answers become more common, businesses will have to judge their communications across several distinct outcomes: whether an announcement was published, whether its intended audience encountered it, whether readers engaged with it, and whether the content served a meaningful business objective.
Those measures are related, but they are not the same. A live article confirms publication. Referral data may show that people visited a company’s website. Inquiries can signal audience interest, while sales outcomes need further analysis to establish what influenced a decision.
The next few years will test whether businesses can move past counting placements and build a clearer understanding of what their communications achieve. In an information environment shaped by both publishers and AI systems, the strongest measure of a public announcement may not be where it appears, but whether it gives people accurate, relevant information they can use to make an informed judgment.