Do AI Overview query bubbles send traffic to AI Mode?
Quick answer: Yes. In a current test, Google shows query bubbles at the top of AI Overviews in some languages. Tapping one opens AI Mode with the bubble's text as the prompt. It doesn't refine the regular results. That follow-up is a new query with its own answer and its own citations, so businesses need content that answers the next likely question as well as the first one.

Put your organic marketing on autopilot
artiql researches, writes and publishes SEO + GEO content in every language — and turns each article into a video. See it run on your brand.
What are AI Overview query bubbles and what do they do?
AI Overview query bubbles are clickable suggested prompts. Google is testing them at the top of AI Overviews, and tapping one sends the searcher into AI Mode with that bubble's text as the new prompt. The test was spotted in late October 2026 and appears limited to some languages for now. What caught search watchers off guard is that the bubbles look like the old refinement chips but behave differently. A refinement chip reshapes the classic results page around a narrower query. These bubbles skip the results page completely. The searcher leaves the summary and lands in a chat-style AI Mode conversation that already has their next question typed in. So the bubbles do send traffic to AI Mode. Whether any of it reaches your website is a separate question.
The test fits a pattern Google has been building for about a year. On mobile, tapping "Show more" on some AI Overviews has been tested as a direct route into AI Mode. Google later confirmed that AI Overviews may expand automatically for some queries, which removes that tap and pushes the traditional blue links further down the page. More recently, an expanded AI Overview panel inside Knowledge Panels was spotted with a search box that redirects to AI Mode. Each test is small. Taken together, they show Google building more and more ways into AI Mode from the regular results page. Query bubbles may be the most frictionless of these so far, because the searcher doesn't have to type anything.
For a business owner, the honest reading is that clicks from AI Overviews will probably keep getting harder to win. Every searcher who goes from the summary into a conversation is one who didn't scroll down to the ten blue links. Visibility doesn't vanish, though. It moves into the sources AI Mode cites, the cards it shows and the links it puts in the answer. Google has been testing how those links look, and our breakdown of bigger, bolder website links in AI Mode covers what that could mean for clicks. The practical shift is from "Do we rank?" to "Are we cited when the conversation keeps going?"
Why is the follow-up prompt a separate citation opportunity?
The follow-up prompt counts as a separate citation opportunity because AI Mode treats the bubble text as a brand-new query. It builds a new answer and picks new sources. Being cited in the AI Overview doesn't guarantee you'll be cited in the AI Mode response that follows, and a page that missed the first answer can still win the second. Many site owners assume their AI visibility is one slot per search. With query bubbles, a single search can produce two answers in a row, often on different angles of the same topic. Each one is a chance to be named, linked and remembered, and each one is a chance to lose that spot to a competitor who covered the angle better.
Here's how that plays out for a dental clinic. Someone searches "how much does teeth whitening cost" and gets an AI Overview summarising typical price ranges. The bubbles above it might suggest follow-ups such as whether whitening is safe for sensitive teeth, how long results last, or how in-office whitening compares with at-home kits. Each of those becomes an AI Mode prompt. A clinic with one strong pricing page could be cited in the Overview and then vanish from the follow-up, because nothing on its site answers the sensitivity question directly. A rival clinic with a clear page on sensitive teeth gets the second citation, and that's the one the searcher reads as they move closer to booking.
Our view, and it's an unpopular one: the second step matters more than the first for most businesses. The opening AI Overview usually answers a broad, early question. It's useful for awareness but rarely the moment someone decides. The follow-up is where the question gets sharper and closer to a buying decision: safety, timing, comparisons, "near me", "for my situation". Teams that pour all their effort into winning the Overview are working on the less valuable half. If you have to pick a place to start, build answers for the questions people ask next. They're more specific, there's usually less competition, and they're much closer to the moment someone becomes a lead.
| Step 1: AI Overview | Step 2: AI Mode follow-up | |
|---|---|---|
| What triggers it | The searcher's original query | The text of the bubble they tapped |
| Where it appears | Top of the regular results page | A separate conversational AI Mode view |
| Typical intent | Broad, early research | Narrower, closer to a decision |
| Sources | Chosen for the original query | Chosen again for the new prompt |
| What wins the citation | A clear answer to the main question | A clear answer to the next question |
How does query fan-out shape which pages get cited in AI Mode?
AI Mode uses a technique Google calls query fan-out. It breaks a prompt into related subtopics, runs many searches at the same time, and combines what it finds into one answer. A single follow-up prompt can therefore pull sources from pages that rank for a dozen related queries, not just the exact wording in the bubble. This is why query fan-out matters for SEO. Ranking for one keyword is no longer the whole game. If the bubble says "is in-office whitening better than at-home kits", the fan-out might also look up costs, sensitivity, how long results last and dentist recommendations. Pages that cover those subtopics clearly and accurately are the pool AI Mode chooses from.
In practice, depth across a topic beats one sprawling page stuffed with a single keyword. A tight set of pages, each answering one question well and linking to the others, gives the fan-out more clean entry points to your site. Eligibility still comes first. Google has said that pages need to be indexed and eligible for snippets to appear in its AI features. Beyond that, no special markup or file is required. Some site owners have used snippet controls to limit how Google quotes them, which can quietly take them out of the citation pool. Before you change those settings, read our explainer on what nosnippet and Google's AI controls really do to your visibility in AI Overviews.
Fan-out also explains why AI Mode citations can feel unpredictable from week to week. The sub-searches behind an answer change with the wording of the prompt, the searcher's context and the documents Google retrieves at that moment. You can't control all of it, but you can raise your odds by being a strong match for many of the subqueries a topic usually triggers. More AI Mode traffic may also come from outside the search page. Google has been working on a Chrome feature that decides whether a query goes to AI Mode or to classic results, and our piece on how Chrome may route searches to AI Mode explains why citations now matter as much as rankings.

Put your organic marketing on autopilot
artiql researches, writes and publishes SEO + GEO content in every language — and turns each article into a video. See it run on your brand.
How should you structure content around the next likely questions?
Structure content for query bubbles in three ways. Write headings as the questions people actually ask. Open every section with a direct, quotable answer. Organise pages into topic clusters that cover the follow-up questions around each core subject. That way your site can be cited for the opening query and again when the searcher taps a bubble. Question-style headings help both readers and retrieval systems match a passage to a prompt. An answer-first paragraph gives AI Mode a clean sentence to cite on its own. Clusters make sure there's a relevant page waiting when the conversation turns to cost, safety, timing or comparisons. None of this is a trick. It's what clear writing looks like when a machine is reading over the visitor's shoulder.
Finding the next likely questions takes less guesswork than you might expect. Start with the bubbles and follow-up suggestions Google already shows for your main queries. Search them yourself on mobile and in AI Mode, and note every angle that comes up. Add the People Also Ask boxes, the queries in your Search Console performance report, and the questions customers really ask on sales calls, in support tickets and at the front desk. That last source is often the richest, and competitors can't easily copy it. Group everything by the main topic it belongs to, then by the stage of the decision it reflects. Each group becomes a cluster: one central page with supporting pages around it, each answering a single follow-up clearly.
The writing inside each page matters just as much. Put the answer in the first one or two sentences, then add the evidence, examples, trade-offs and edge cases a careful reader would want. Use specific numbers only when you can back them up. Name the situation where your advice doesn't apply, because careful qualifications make a page more trustworthy to people and more useful as a source. Keep each section focused on one question. Long paragraphs that cover three ideas are harder to quote. To check how ready your existing pages are, run them through a free GEO score check. It shows where structure, clarity or answer-first writing is holding a page back from being cited.
How do you measure whether you're cited at both steps?
You can't yet see query bubble clicks separately. You can still measure whether you're cited at both steps by combining Search Console data with regular manual checks of AI Overviews and AI Mode for a fixed set of prompts. Google includes AI Mode activity in Search Console's performance report under web search, so clicks and impressions are counted but not broken out by feature. That makes your own tracking essential. Make a list of 20 to 40 prompts: your main queries plus the follow-ups the bubbles suggest. Each month, check whether your site appears in the AI Overview, in the AI Mode answer, or in neither. A simple spreadsheet is enough. What matters is checking the same prompts every month, not how fancy the tool is.
Next, connect those citation checks to the outcomes that pay the bills. Watch for rising impressions on long, conversational queries in Search Console, because that's often the first sign that AI Mode is pulling from your pages. In your analytics, note branded searches and direct visits that rise after a new cluster goes live. People who read your name in an AI answer often come back later on their own. Google has also been testing extra link placements in AI Mode, and our guide to the Further exploration section shows how to become eligible. If you want a rough forecast of what better visibility is worth, the SEO traffic calculator turns target rankings into estimated visits.
There's a clear business case for doing this work instead of buying your way around it. Paid ads can put you above an AI Overview, but they don't earn a citation inside the AI Mode conversation that follows. That traffic also stops the day the budget runs out. A well-built topic cluster keeps answering follow-up questions for months after it's published, and every new supporting page strengthens the rest. If you're weighing the two, our 12-month comparison of SEO and Google Ads costs shows when organic content starts to beat ad spend. Artiql builds these answer-first clusters for you on autopilot. Pricing is set for each business after a short discovery call, so the best next step is to book a demo and see your own prompts mapped out.
Frequently asked questions
Are AI Overview query bubbles available to everyone?
Not yet. Query bubbles are a Google test, seen so far in some languages, and Google hasn't announced a full rollout. Tests like this often change or disappear. Still, the direction is consistent: Google keeps adding ways to move searchers from AI Overviews into AI Mode. Building content for the next likely question pays off whether or not this exact feature launches widely.
Do I need special markup to get cited in AI Mode?
No. Google has said there's no special markup, file or tag needed to appear in its AI features. Your pages need to be indexed and eligible to show a snippet, the same basics that classic search requires. After that, clear question headings, direct answers in the opening sentences, accurate details and solid internal links within topic clusters do most of the work.
Should I block AI Overviews so searchers have to click through?
For most businesses, no. Blocking or limiting snippets can take your pages out of the pool AI Overviews and AI Mode cite from. You won't get clicks back, and competitors will take the visibility you gave up. Blocking can make sense for paywalled or licensed content you can't afford to have summarised. For a clinic, firm or online store, being cited usually helps more than staying hidden.
How long does it take new content to get cited in AI Mode?
There's no fixed timeline. A new page has to be crawled, indexed and judged relevant before AI Mode can use it, which can take days on a site Google trusts or several weeks on a newer one. Pages that link into an existing topic cluster usually get picked up faster. Track a set list of prompts monthly rather than expecting results overnight.

Put your organic marketing on autopilot
artiql researches, writes and publishes SEO + GEO content in every language — and turns each article into a video. See it run on your brand.