A small study found the click-through rate from Google's AI features, AI Overviews and AI Mode, is about 0.9%, and when a site opted out of showing in Google's generative AI features, it also saw a drop in clicks in the main web results. The result captures the opt-out dilemma in miniature: leaving AI answers may feel like protecting your content, but in this test it cost the site visibility and clicks where it mattered most, in the traditional results themselves.
The test's two findings need to be read together. First, the click-through rate from Google's AI features, AI Overviews and AI Mode, is about 0.9%. That is a low rate, and it confirms what many publishers have suspected: appearing inside an AI answer does not guarantee a click, because the answer often satisfies the reader without one. Second, when a site opted out of showing in Google's generative AI features, it also saw a drop in clicks in the main web results. That second finding is the surprising one. Opting out was expected to affect AI visibility; instead, it was associated with fewer clicks in the ordinary organic results as well. For a site considering an opt-out, that combination is difficult to ignore.
A click-through rate of about 0.9% from AI features sounds small, and in isolation it is. But it should not be read as proof that AI visibility is worthless. Clicks are only one measure of value: an AI citation can build awareness, establish authority, and put a brand in front of readers who later search for it by name or visit directly. At the same time, a low click-through rate does explain publisher frustration. If a site's content helps produce an answer that thousands of people read, but only a small fraction click through, the site bears the cost of creating the content while capturing little of the attention it generates. That mismatch between contribution and reward is at the heart of the current debate over AI search.
The drop in main web results clicks after opting out is the finding that deserves the most careful interpretation. It does not necessarily prove that opting out directly causes a ranking penalty, and a small test cannot settle that question on its own. But it does suggest that AI visibility and organic visibility are not as separate as publishers might hope. Appearing in both AI answers and traditional organic results could build trust and increase clicks, as the coverage of the test suggests: a reader who sees a brand in an AI answer and again in the organic results may be more likely to click, recognising a name they have just encountered. Removing the AI presence may therefore remove a reinforcement effect, leaving the organic listing to work alone. That hypothesis fits the data, even if the test alone cannot prove it.
For SEOs, the test frames a genuine dilemma rather than a simple recommendation. Opting out may protect a principle, but it may also reduce overall clicks, including clicks that have nothing to do with AI. Staying in may preserve clicks and build familiarity, but it means accepting a low click-through rate from AI features and supplying content that often answers the query without a visit. The right choice will depend on the site, its traffic mix, and its business model, and it should be made as an experiment, not a permanent stance: opt out or opt in for a defined period, measure clicks, revenue, and brand search, and then decide on evidence. A small test cannot dictate strategy for every site, but it can warn against assuming that opting out is cost-free.
Rather than guessing, SEOs should build measurement that separates AI visibility from organic performance, so that any change can be attributed correctly. Track impressions and clicks from AI features separately from traditional results, watch brand search and direct traffic for signs that AI citations are building familiarity, and compare periods before and after any opt-out decision. Content strategy should also adapt: pages that answer questions completely, cite sources, and offer something beyond the answer, such as original data, tools, or expert judgment, give readers a reason to click even when an AI summary exists. In an AI-mediated search landscape, the click has to be earned twice: once by being cited, and again by offering something the summary does not.
Sources: Search Engine Land, October 29, 2026, plus Search Engine Roundtable, October 30, 2026.
Linkbaits helps you create content that earns citations and backlinks. Start with the free tools and build your first linkable asset this week.
Start free trial