Google’s VP of Search on What Publishers Need to Do to Stay Visible in AI Search

Google’s VP of Search, Liz Reid, gave an interview recently in which she addressed the question publishers have been asking for the past 18 months: what do you need to do differently to remain visible in AI search?

Her answer was direct and, for many publishers, challenging to hear. The path to AI visibility is the same as the path to reader engagement. Create content people actually want to read.

Publisher Traffic Loss Is Broader Than AI

Before addressing what publishers should do, Reid made a point that gets lost in most discussions about traffic decline: AI Overviews are not the only cause.

She pointed to a shift in how people consume content, noting that audiences are moving toward video formats and social media for content they would previously have found through search. A Reuters study on changing content consumption behaviour supports this observation, showing that the shift away from text-based publisher content is a broader audience trend, not purely an algorithmic one.

This distinction matters for how publishers diagnose and respond to traffic changes. A publisher experiencing declining referral traffic from search may be dealing with an AI visibility problem, a format relevance problem, or both simultaneously. Treating them as the same issue leads to the wrong response.

What Google Is Actually Looking For

Reid’s description of what earns visibility in AI search is consistent with what Google has been saying through its documentation, but the interview made the standard more concrete.

She described the content that performs well as content that reflects genuine expertise, carries a distinct perspective, and offers something the reader cannot find in the same form elsewhere. In her words, publishers need to produce content that is “not the 1,000th copy of the same story” but rather something with an interesting take on it.

The implicit standard here is non-commodity content. Content that any writer could produce from a search and a basic familiarity with a topic does not meet it. Content that draws on real experience, original analysis, or a perspective specific to the author or organisation does.

Reid said the more publishers build content that their audience will love, the more it will work. The more content is designed for the search engine rather than the audience, the less effective it becomes over time.

The Two Conditions Reid Named

When asked directly what publishers can do to maximise their visibility in AI search, Reid outlined two conditions.

The first is technical access. She said to make content crawlable: if you block the content, that will not work, and if it makes it hard to discover, then that’s difficult. Google’s Search Console provides controls for publishers to manage what is and is not accessible. Making content available to crawl is the prerequisite for everything else.

The second is editorial quality. Reid described the content that works as content that brings in expertise, is fresh and relevant, reflects what people are curious about, and carries experience, detail, and richness.

These are not new criteria. They map directly to E-E-A-T: Experience, Expertise, Authoritativeness, and Trustworthiness. What the interview clarifies is that these standards now apply not just to how content is ranked in traditional search results, but to whether it is cited in AI-generated answers at all.

The Tension Reid Did Not Address

Reid’s framing places the responsibility for traffic decline squarely on publishers. Create better content and the visibility will follow.

That framing leaves a question open that many publishers are sitting with: what happens to high-quality content on topics where AI Overviews resolve the query completely, regardless of how well written the underlying sources are?

A publisher producing genuinely excellent informational content on a topic where Google’s AI can construct a comprehensive answer from multiple sources may still see their click-through rate fall, even if their content is cited in the overview. Liz Reid’s advice addresses the citation problem. It does not fully address the click problem, which the SparkToro zero-click study published earlier this week put at 68% of US searches ending without a click to any external site.

Both problems are real. They require different responses.

What This Means for Content Strategy

The practical reading of Reid’s comments is a straightforward filter for content planning decisions.

Before commissioning or creating a piece of content, the relevant question is whether this piece brings something to the subject that does not already exist in the top ten results. If the answer is a structured summary of publicly available information, that content will not perform in AI search regardless of its technical optimisation. If the answer involves original data, first-hand experience, expert analysis, or a perspective specific to the brand’s position in its market, that content has the properties Reid described as necessary for AI visibility.

Content strategy in the current environment is increasingly a question of editorial positioning, not keyword coverage. The brands that understand that distinction are building a durable advantage. The ones still optimising primarily for search engines rather than readers are building a diminishing one.