Google is testing a stripped-down header in AI Mode that shows only the AI Mode tab, with the other tabs removed from the line, Search Engine Watch reported on November 2. The test follows a recent experiment in which AI Mode tucked the other tabs behind a dropdown under the All tab. Now Google appears to have gone a step further, removing them from the header altogether. Google has not said why it is testing this, and the change should be read as an interface experiment, not as a confirmed redesign.
In classic Google Search, the header shows tabs such as All, Images, Videos, News, Shopping, and more, letting a user switch between result types with one click. In AI Mode, the header has been evolving toward fewer options. First, the other tabs were tucked behind a dropdown under All, requiring an extra click to reach them. Now, in the latest test, the header shows only the AI Mode tab, and the other tabs are gone from the line entirely. A user in AI Mode who wants a different result type must find another route, or leave AI Mode altogether.
Search Engine Watch's reading of the test is that Google does not want users to switch away from AI Mode to other tabs. That is the publication's interpretation, not Google's stated reason, and it should be treated as such. Google tests interface changes for many reasons, including simplification, measurement, and user behaviour, and a stripped-down header could reflect any of those. What is observable is the direction: each successive test makes it harder, or at least less obvious, how to leave AI Mode for another tab.
Headers are not decoration. They are the map of a product, telling users what is available and how to get there. A tab that is visible invites a click. A tab hidden behind a dropdown invites a click only from users who already know it is there. A tab that is removed invites no click at all. Over time, those small frictions reshape behaviour: users who cannot easily switch tabs will stay in the surface they are in, and will learn to use that surface for tasks they once did elsewhere.
For AI Mode, that has a specific implication. If users stay in AI Mode because switching is harder, then AI Mode becomes the default search experience for those users, not just an option. Their queries, their clicks, and their expectations will all be shaped by AI Mode's answers, citations, and ads, rather than by classic results. That is a powerful outcome for a header change, which is why interface tests deserve attention even when they look trivial. The header is where a product's strategy becomes user behaviour.
For SEOs, the test reinforces that AI Mode is not a side feature. It is becoming the main surface for a growing share of users, and the tabs that once let users escape to classic results, images, or news are being de-emphasised. Content that only performs in a specific tab, such as an image or news result, may lose visibility if users stay in AI Mode and never switch. Content that performs inside AI Mode answers, through citations and clear, quotable information, will gain relative importance as the header narrows.
For publishers, the implication is to optimise for the answer, not for the tab. A news publisher that relies on the News tab, or an image publisher that relies on the Images tab, should check how its content appears inside AI Mode, and whether AI Mode cites it for relevant queries. If it does not, then a header test that keeps users in AI Mode could reduce that publisher's traffic even without any ranking change. The defence is the same as for other AI search shifts: be citable, be accurate, and be present in the answer itself.
The next evidence will be whether the stripped-down header spreads beyond the test group, and whether Google comments on its purpose. If it ships widely, it will confirm that Google wants AI Mode to be a self-contained experience, not a gateway to other tabs. If it disappears, it will join the many interface tests that taught Google something and then vanished. Either way, the sequence of tests, from dropdown to removal, suggests a direction of travel that is worth tracking, even if this particular version never ships.
SEOs should also watch the companion tests reported in the same period, including new follow-up queries at the bottom of AI Mode answers and a Where to learn more heading above citation cards. Together, those tests describe an AI Mode that is trying to keep users inside the conversation, by offering follow-ups, citations, and answers, while making it harder to leave for another surface. That is a coherent product direction, and it is one that publishers and SEOs should plan for, whether or not any single test ships.
Sources: Search Engine Watch, November 2, 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