AI Overviews: More Links, But Which Clicks Actually Reach Your Site?
Summarize this article with AI
Avant d’entrer dans le dΓ©tail, posons le mΓ©canisme. Une seule interaction dans un AI Overview, deux issues possibles : votre site, ou une nouvelle requΓͺte qui reste chez Google.
Un lien, deux destinations
Dans un AI Overview, un mΓͺme lien mΓ¨ne soit Γ votre page, soit Γ AI Mode.
An AI Overview Link Is Not Always a Click to Your Site
One link. Two destinations. One stream of traffic.
I look at a client’s e-commerce analytics. He sells hiking gear in Southeast Asia. His product pages get cited in Google’s generated answers. Sessions lag behind. Search Console shows stable impressions. The site loses visitors. The contradiction is obvious.
Visibility is not the issue. The link is.
In an AI Overview, a link can lead to a regular web page. That’s the one you care about. But it can also open a conversation in AI Mode. Google announced it on January 27, 2026: you can ask a follow-up question from an AI Overview and switch into an exchange with AI Mode. The link does not leave Google. It triggers a new search.
Search Engine Journal published an analysis this week. The finding: AI Overviews have more links than before, but not all of them return to the web. On September 21, 2026, Gagan Ghotra showed a screenshot where underlined text in an AI Overview opens an AI Mode conversation instead of a site. Nothing abnormal. It’s one of the two intended destinations.
Result for an e-commerce merchant: you’re cited. Your page is visible. But the click doesn’t belong to you. Measuring your organic traffic becomes fuzzy.
What exactly are you tracking?
On August 25, 2025, Robby Stein, vice president of product at Google Search, announced the addition of link carousels on desktop. Goal: more inline links. Since then, Google has piled up touchpoints. On February 17, 2026, link groups appear on hover on desktop. On May 6, 2026, Hema Budaraju details five updates, including reading suggestions at the end of answers and previews of public discussions.
These changes transform your reporting. Each touchpoint has its own click logic.
There’s the link that leads to your product page. The one that opens a conversation. The one that shows a preview. The one that cites a Reddit discussion. And the one that stays inside Google.
Only one of these paths drives revenue.
For a mobile shopper, the difference is invisible. They tap an underlined word, expect a product page, and land on a new generated answer. Frustration sets in. But Google’s idea is clear: keep the user in the conversation.
Search Console Counts One Click, Not the Other
Zero impressions. Zero clicks.
When a link in an AI Overview triggers a new Google query, Search Console records it neither as an impression nor as a click. That’s Google’s rule, echoed by Search Engine Journal. A click to an external page counts as a regular click. Simple to understand. Heavy with consequences.
For an e-commerce merchant, the trap is precise. You see a query appearing in your reports. You see impressions. You believe your pages capture attention. But half of the possible interactions disappears from the count. Not because you’re doing poorly. Because Google doesn’t trace them as web traffic.
The table below summarizes the mechanics.
| Link display | Destination | Search Console count |
|---|---|---|
| Underlined text in the answer | External web page | Click + impression |
| Desktop link carousel | External web page | Click + impression |
| Mobile follow-up question | AI Mode (Google query) | None |
| Public discussion preview | Reddit or forum thread | Click to third-party site if exit |
| End-of-answer suggestion | Article or web page | Click + impression |
This table doesn’t come from an internal bulletin. It derives from Google announcements between August 2025 and May 2026. The nuance is in the destination, not the link.
Consequence for your dashboards: you’re comparing things that aren’t comparable. A rise in AI Overview impressions doesn’t mean a rise in sessions. A drop in sessions doesn’t necessarily indicate a loss of visibility. Clarity comes from separating the signals.
What you measure changes your decisions. If you invest in a page cited in AI Mode without web tracking, you have zero data to judge the impact. Zero. A measurement black hole.
In practice, two mistakes recur among e-commerce merchants. The first: panicking at a drop in sessions and deleting pages cited in AI Overviews. The second: reallocating budget to SEO believing AI Overviews bring nothing, when citations build lasting presence. Same misunderstanding.
The click to your site is one line. The passage through AI Mode is another. They don’t carry the same commercial weight. But both need to be read with the right tool.
Add hover link previews. On February 17, 2026, Google announced link groups with descriptive icons. On desktop, the user hovers, reads the site name, then decides. This reading time isn’t measured. It can transform a click into a non-click. Search Console stays silent on these micro-decisions.
For a catalog of 800 references, each friction point repeats hundreds of times per month. The gap between impressions and clicks becomes a confidence indicator.
Google ne cache pas sa stratΓ©gie. Cette frise reprend les annonces qui ont faΓ§onnΓ© les liens dans les AI Overviews, d’aoΓ»t 2025 Γ mai 2026.
La chronologie des annonces Google
De Robby Stein Γ Hema Budaraju : comment Google a multipliΓ© les liens.
Why Google Multiplies Links Without Sending You the Click
The answer fits in one sentence: retain the user.
Google doesn’t hide its intentions. On August 25, 2025, Robby Stein announced that the main goal of AI Mode was to display more inline links. On February 17, 2026, he described link groups on hover on desktop, with descriptive icons. On May 6, 2026, Hema Budaraju listed five updates, including reading suggestions at the end of answers and previews of public discussions. Every announcement points in the same direction: multiply exit points, but not all toward the web.
This strategy has a logic. An inline link to your site can pull the user out of Google. A link to AI Mode keeps them inside Google. A Reddit discussion preview keeps them in the social ecosystem. An article suggestion at the end of an answer keeps them in Search, often toward another piece of content. The more the user clicks, the longer they stay.
- August 25, 2025: desktop link carousels.
- January 27, 2026: mobile follow-up question to AI Mode.
- February 17, 2026: hover link groups with icons.
- May 6, 2026: end-of-answer suggestions, discussion previews, subscription labels.
For an e-commerce merchant, it’s both a boon and a trap. The boon: AI Overviews cite more sources, so more product pages can appear. The trap: some of these citations generate zero measurable sessions. You’re feeding a storefront that doesn’t always send you the shopper.
Client pages show rising impressions. Clicks stagnate or drop. The click-to-impression ratio erodes. Content has nothing to do with it. The architecture of the Google ecosystem absorbs the gesture.
In May 2026, Google clarified that hovering an inline link shows a preview with the site name or page title. This detail reassures the user before the click. But it adds a step. How many users hover, read the preview, then stay in Google? Impossible to trace in Search Console. A gray measurement zone.
This mechanism pushes us to rethink the success metric. The click remains important. But the citation, the impression, and presence in previews also count. Especially on mobile, where tracking is even fuzzier.
Everything works as designed. The paradigm has changed.
On September 21, 2026, Peec AI shared a chart showing the percentage of AI Overviews containing an external link in the answer text. The trend climbs over the period from August 25 to September 20. Not an isolated spike: a continuous progression.
For you, e-commerce merchant, this means your product pages have more chances of being cited than a year ago. But also more chances of being cited in a format that doesn’t send the click back to you. The math gets trickier.
Regardons le cas terrain de prΓ¨s. Sans pΓ©nalitΓ© ni contenu obsolΓ¨te, ce marchand d’Asie du Sud-Est voit ses sessions fondre : voici la dΓ©composition du delta.
OΓΉ sont passΓ©es les 1 300 sessions ?
Chez ce marchand, le trafic organique glisse de 14 200 Γ 12 900 sessions.
What I Observe at an E-commerce Client: Sessions Slide, Impressions Hold
A Southeast Asian e-commerce merchant sells small office furniture. Catalog: 847 references. Before the expanded AI Overviews rollout, its monthly organic traffic hovered around 14,200 sessions. Product pages captured 68% of those visits. Listings were clean. Customer reviews plentiful. Nothing foreshadowed the break.
In six weeks, product sessions lost 9%. Search Console shows 14,200 sessions before, 12,900 after. That’s -1,300 sessions per month. Not a collapse. A quiet erosion.
First reaction: look for a penalty. None. Second: tell yourself the content was outdated. False. The pages kept appearing in regular results. Positions hadn’t moved on major queries.
The signal was elsewhere: in impressions from AI Overviews. They were climbing. Clicks weren’t. The click-through rate on these impressions fell from 4.2% to 1.8%. One detail: these numbers only count interactions where Google traces a click to the site. Links to AI Mode appear nowhere.
On product pages: 23,400 AI Overview impressions before the shift, 31,200 after. Clicks dropped from 980 to 560. That’s -43%. No penalty, no loss of regular ranking. Just a shift in user behavior.
I cross-referenced the data with Google announcements. On January 27, 2026, the mobile follow-up question arrived. On May 6, 2026, reading suggestions and discussion previews were added. The chronology lines up. Session erosion starts when Google multiplies internal touchpoints.
Case summary: the site wasn’t losing visibility. It was losing clicks to the Google ecosystem. Product pages remained cited. But the mobile user tapped the underlined text and opened AI Mode. Or hovered the preview and turned back. Or checked a Reddit discussion. Direct traffic dried up with no fault from the publisher.
The pattern isn’t isolated. I see it on other catalogs. The gap between impressions and clicks widens. E-commerce merchants who only look at sessions wrongly conclude that SEO no longer works.
The uncomfortable reality: SEO works, but it also fuels Google’s retention. Full clarity.
Rethink Your KPIs Before You Panic
Click less. Measure better.
Start by revising your definition of organic traffic. A session that lands on a product page remains a strong signal. An impression in an AI Overview, a link preview, a citation in a public discussion: that’s weaker. Useful, but not the same thing.
In Search Console, separate the data. The Performance tab gives a breakdown by search type. Filter on impressions that come from AI Overviews. Compare clicks, not just impressions. The delta shows you the share Google absorbs.
Add citation tracking. Google doesn’t provide a complete report on links to AI Mode. But you can monitor cited pages with monitoring tools. A product page cited in a generated answer? Note it. Over 60 days, you’ll see if it generates direct sessions, brand searches, or nothing.
Next, keep your cited pages, even if they don’t convert to a direct click. AI Overviews build presence. A shopper who sees your brand cited, then comes back later via a brand search, has been influenced. This journey doesn’t appear in the immediate click report. It plays out in awareness, not in sessions.
E-commerce merchants should track three indicators together:
- Clicks to the site from AI Overviews.
- AI Overview impressions on product pages.
- Brand traffic and navigational searches over 30 days.
If the first two move in opposite directions and the third climbs, your AI Overview presence is doing its job. If all three drop, then there’s a deeper problem.
I see this in the field. Clients who accept this reading grid make better content decisions. They stop chasing phantom sessions and invest in pages that capture attention, even without a click.
And if you sell a high-consideration product, the effect is even stronger. The buyer compares, reads, comes back. The direct click is just one step.
The direct click remains the ultimate proof. But the citation is the first step.
Your dashboard should reflect this journey. Not just the finish line.
Otherwise, you’re flying with an old map. One that no longer shows the right roads.
A field tip: segment by device. On mobile, follow-up links to AI Mode are more frequent. On desktop, carousels and previews dominate. Your decisions change depending on your catalog’s share of mobile traffic.
AI Overview Traffic Can No Longer Be Measured in Sessions Alone
One link. Two destinations. One revenue stream.
AI Overviews send traffic. That’s a given. The question remains what share returns to your site and what share stops inside Google. Search Engine Journal just confirmed it: more links, but not all of them return to the web.
In September 2026, a citation can no longer be judged by its click. Nor a product page by its session. Signals are decoupling. It’s new, and it’s uncomfortable. For those who measure correctly, it’s also an opening.
The direct click keeps its value. It signals strong intent. But it no longer covers the entire journey. The rest plays out in citations, previews, discussions. That’s where the decision takes shape.
For an e-commerce merchant, the margin lies in the delta between AI Overview impressions and clicks to the site. If the delta climbs, your pages still serve Google, but not yet you. It’s not about deleting. It’s about spotting which citation format triggers the direct click.
Test, observe, adjust the content of cited pages. A product page with a more complete answer can resurface in end-of-answer suggestions. A clear snippet can lift the click-through rate to the site. The lever is there.
I don’t sell a method. I show pages. Open your Search Console. Filter on AI Overview. Compare clicks to impressions. That simple look changes the conversation.
Direct traffic doesn’t disappear. It transforms. Those who measure citation and session separately get ahead. Those who wait for a return to the old world decide with amputated data.
Attribution works for you. Not against you.
And you, how many of your AI Overview links bring a visitor back to your catalog?
SEO Audit: Measure Your AI Overview Citations
I look at your cited pages, your direct clicks, and your AI Overview impressions. You leave with a clear read on what really comes back to your site.
Book a strategic call β 45 minFrequently Asked Questions
How do I know if a click comes from an AI Overview or a regular result?
In Search Console, filter performance by search type. AI Overview impressions and clicks appear on two separate lines. If the link opens AI Mode, the data stays at zero.
Why are my sessions dropping while my impressions remain stable?
Your pages appear in AI Overviews. Except some links go to AI Mode, not your site. And those clicks, Search Console doesn’t see them. So look at two numbers: AI Overview impressions and your actual clicks.
Does a citation in AI Overview without a click have value?
Yes. It plants the brand in people’s minds, and it resurfaces later in brand searches. Watch brand traffic over 30 days: that’s where you see the effect.
Should I delete cited pages that no longer generate direct clicks?
No. They create visibility in Google. Before deciding anything, check if they generate AI Overview impressions and brand traffic.
How do I track links that lead to AI Mode?
Google releases no report on this. I use monitoring tools to spot cited pages, then cross-reference with Search Console and my analytics.

