Google is testing dropping the citations card from AI Overviews. The citations card is the panel that surfaces source links alongside an AI-generated answer, and its removal in the test changes how prominently cited sources are displayed to searchers.
The citations card collects the sources behind an AI Overview into a visible panel, giving readers a path from the summary to the underlying pages. For publishers, it is one of the main routes by which an AI Overview converts into a visit: a reader who wants more detail clicks through from the card.
Removing the card does not necessarily remove citations entirely; links may still appear inline within the answer text. But it changes the visual weight of sourcing, and a less prominent citation path typically means fewer clicks, even if the same links technically remain.
Google constantly tests answer layouts to balance completeness, readability, and monetisation. A citations card takes vertical space and attention, and on mobile that space is expensive. If inline links perform as well as a card for user satisfaction, dropping the card simplifies the answer.
The test also fits Google's October 15 direction of making links more prominent in some places, with higher placement and thumbnails, while simplifying elsewhere. The two moves are not contradictory: Google is experimenting with which link presentation actually drives engagement, rather than assuming one layout fits every query.
Publishers cannot opt into or out of a layout test, but they can watch its effects. If a test like this rolls out widely, referral traffic from AI Overviews would be expected to fall for cited pages, even if citation frequency stays the same, because the click path gets weaker.
The metric to watch is Search Console's AI-feature impression data alongside analytics referrals from Google search. Impressions without clicks would confirm that citations are present but less actionable, which is a different problem from losing citations altogether, and it calls for different advocacy.
Layout tests come and go, but the underlying requirement does not change: pages that are clear, authoritative, and directly answer the query are the pages AI systems cite. Structured headings, direct answers near the top, and original data all increase the chance of citation in any layout.
Also make sure cited pages reward the click. A reader who does click through from an AI answer should land on a page that expands the answer, not a homepage or a thin doorway page. Pages that disappoint clickers teach both readers and systems not to cite them again.
Citation visibility is becoming a ranking-adjacent asset: being cited inside an AI answer is valuable even before the click, because it builds brand familiarity and because other writers see and link to cited sources. A weaker citations card reduces the click value of a citation but not necessarily its brand value.
Track both. Count citations and brand mentions in AI answers separately from referral clicks, so a layout test that cuts clicks does not look like a total loss of visibility, and so you can argue for the value of citation work with data rather than anecdotes.
Whether inline links or a citations card drives more clicks is an empirical question, and Google's test is designed to answer it. Inline links have the advantage of context: the reader clicks the exact claim they want to explore. Cards have the advantage of completeness: the reader sees every source in one place and can choose.
For publishers, the practical difference is small compared with the bigger question of whether they are cited at all. A page that is cited inline and a page that is cited in a card both gain brand exposure, and both depend on the same underlying quality. Optimise for being cited first, and worry about the layout second.
Sources: Search Engine Roundtable podcast recap, October 29, 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