The shift is mostly invisible from a comms team's day-to-day, but the data is unambiguous. Google's AI Overviews, which expanded broadly in 2024 and have continued to roll out across query categories, now sit at the top of an enormous share of search result pages. For informational queries — the kind that drive a lot of organic discovery for brands and publishers — AI Overviews intercept the click before it ever happens.
The publishing industry has noticed. The Reuters Institute's 2026 trends report (here) documents news organizations forecasting steep declines in search referral traffic over the next three years, with many citing AI Overviews and similar zero-click surfaces as the proximate cause. For communications teams, the question is whether the traditional value of earned media coverage — and the press releases that support it — survives the shift.
What an AI Overview is doing under the hood
When a user submits a query that Google decides to handle with an AI Overview, the system retrieves a set of sources, synthesizes a paragraph-or-two answer, and surfaces the answer above the traditional results. Cited sources appear inline as small icons or links, with the option to expand. The user can scroll past the Overview to the traditional results — and most users do see those results — but the click-through rate to source content has been declining.
This is the structural shift earned media has to plan around. The old model assumed that a press release or earned story would generate clicks to the publisher, where readers would absorb the brand's message. The new model assumes the brand's message has to land in the synthesized answer itself. The publisher's authority still matters — Google is selecting which sources to cite — but the brand's exposure now happens in the AI surface, not on the publisher's page.
Implications for press release strategy
The press release as a format has been declared dead approximately every 18 months for the last decade. It keeps not dying because it serves a function nothing else does: it is structured, dated, machine-readable content that puts a brand's facts on the record in a form retrievable systems can parse.
The strategic adjustment is in how press releases are written. Releases optimized for AI surfaces front-load the substantive claim, include clear context, name the relevant entities (people, products, companies), and avoid the hype-language that retrieval systems tend to discount. A release that reads like marketing copy is less likely to be cited than one that reads like a wire service news brief. This convergence — where the most retrievable releases look most like journalism — is healthy. It rewards substance.
Implications for earned media targeting
If the placement's value is partly determined by whether it ends up cited in an AI Overview, the targeting calculus shifts. A few practical effects.
First, editorial credibility is repriced upward. Outlets with established authority — major dailies, recognized trade publications, government and academic sources — get cited disproportionately. Pitching strategy that prioritizes reach over authority is increasingly inefficient.
Second, freshness matters more for time-sensitive categories. Google's AI Overviews surface recent content for queries where recency is relevant. A six-month-old feature carries less retrieval weight than a six-week-old one, all else equal.
Third, structured information beats narrative for some query types. Lists, tables, comparison charts, and clear data points get pulled into Overview answers more reliably than long-form narrative passages. PR teams placing data-driven content should ensure the data is presented in a parseable form.
What this is not
It is not the death of search-driven discovery. Google still drives the overwhelming majority of category research traffic across the web. Direct referrals from AI tools remain a small fraction of total publisher traffic, even as they grow rapidly. The shift is real and consequential, but it is not a sudden cliff.
It is also not a reason to abandon owned content investment. Brand newsrooms, executive blogs, and well-structured resource pages remain among the most reliable inputs to retrieval systems. The owned layer feeds the earned layer feeds the AI layer. Each step compounds.
What to do this quarter
Audit current top-of-funnel queries. Check which return AI Overviews and which return traditional results. Document which sources are getting cited in the Overviews for category queries, and which are not.
Then prioritize earned media work toward outlets that already appear in those Overviews. The signal is empirical and available — most teams just are not looking at it yet.
The 10-blue-links interface is not gone, but it is no longer the primary surface. Comms teams that adjust now will compound a structural advantage. The ones that wait for a clearer playbook will be reading about it in a Reuters trends report two years from now.
