AI Overviews and Organic Traffic: More Links, More Visits?

Google is making website links more prominent within its AI-generated answers. For anyone publishing online, that sounds like good news: if links are easier to spot, people should have a better chance of opening them. Yet there is still a decision between noticing a source and visiting its website.

This is where the relationship between AI Overviews and organic traffic becomes interesting. A page can help answer a question, appear as a source and receive no click. Another page might attract fewer visits than before while reaching people closer to making a decision. Those situations need different measurements and different responses.

Starting with Google's announcement on October 15, 2026, I want to explore what could change for businesses, professionals and publishers. The central question is about the practical value of visibility: which content still deserves a visit, and how can we tell whether our SEO work is producing useful results?

What Google announced about links in AI answers

In its October 15 announcement, Google describes more inline links, appearing higher in answers more frequently. It also outlines related-article links with thumbnails and wider availability of news carousels. Tests will highlight official pages and link creator names to Search profiles.

The changes are at different stages: some are being introduced, while others are experiments. The announcement provides no measured increase in publisher clicks or complete market-by-market timetable. It describes a development to watch, rather than a guaranteed recovery in website traffic.

Why AI Overviews and AI Mode should be considered separately

AI Overviews are summaries that can appear within Google's search results. AI Mode offers a more conversational search experience, including complex questions and follow-up exploration. Google's guidance for website owners explains their role.

That distinction matters when planning content. Someone looking up a definition may stop after a short explanation. Someone comparing suppliers, assessing an investment or troubleshooting a technical issue may need supporting evidence and greater detail. Those searches should not be expected to generate the same likelihood of a visit.

For Webita, a general definition of SEO serves a different need from a question about investigating a decline in enquiries from Google. In the second case, the reader may need to interpret figures, compare possible explanations and decide which improvements deserve a budget. Useful depth depends on the problem, rather than on an article's word count.

Mentions, citations and visits are different outcomes

A mention occurs when a brand name appears in an answer. A linked citation gives the reader a route to a source page. A visit happens when someone follows that route and reaches the website. Beyond that lies the business outcome: an enquiry, purchase, subscription or another meaningful action.

These steps can connect, but one does not automatically guarantee the next. An answer naming a business without linking to it may build recognition without an immediately attributable visit. A linked article can also provide enough information within the answer that the reader has no reason to open it.

A citation is a visibility signal. A click is a behaviour. An enquiry is an outcome: I would keep those distinctions in the reporting too.

This does not make a mention worthless. It means being clear about the objective. A publication funded by advertising may assess visits differently from a B2B service that wins a small number of customers through a lengthy decision process.

Will more links in AI answers actually generate more organic traffic?

My reading is that clearer links could remove one obstacle: difficulty finding the source. Another obstacle remains, and it may matter more: the reader needs a reason to open it. If Google's answer already meets the need, a more prominent link may leave the decision unchanged.

A Pew Research Center study published in July 2025 helps illustrate the distinction. Using browsing data from 900 US adults in March 2025, it found traditional-result clicks in 8% of visits to search pages with an AI summary, compared with 15% without one. Links within the summary were clicked in 1% of visits to pages displaying a summary.

Those are historical observations from a US sample. They do not measure the October 2026 changes, automatically describe other markets or establish that AI alone explains every difference. They do show why the presence of a source link should not be treated as equivalent to a website visit.

An interface change can create an opportunity. Establishing whether that opportunity becomes traffic requires observations from your own website, over a sufficient period and across relevant groups of searches.

Evidence from Italian websites adds another perspective

A SEOZoom study published on October 23, 2026 examined 13,098 Italian websites. Comparing June–August 2026 with the same months in 2025, it reported a 12.1% decline in organic clicks, or around 63 million fewer clicks across the quarter. That is a finding for the analysed sample, rather than a forecast for every website.

The figures cannot assign the entire decline to AI Overviews, and they predate October's interface changes. I would take them as a reason to investigate actual page performance, avoiding both alarmism and the assumption that a stable ranking will always deliver the same amount of traffic.

Which content still gives people a reason to visit?

Imagine two articles about CRM software. The first explains that a CRM organises contacts and sales opportunities. The second compares two setups for a business with five salespeople, explains duplicate records and identifies questions to ask before migrating customer data. This is a hypothetical example, but the editorial difference is concrete.

The definition in the first article may be exhausted in a few lines. The second can support a decision that requires context. That does not mean adding detail for its own sake: the information needs to be relevant, verifiable and specific enough to help someone make a choice.

The same applies to a comparison, a documented case, a reusable template or an original analysis. Those resources must actually exist and add value. A download button leading to a PDF that merely repeats the page does not automatically create a compelling reason to leave Google.

I would also avoid withholding the basic answer to force a click. A useful page should provide orientation immediately and then make deeper investigation possible. Its value lies in what the reader can understand or accomplish by continuing.

Technical SEO remains a practical foundation

Google states that foundational SEO still applies, without special AI markup. A page needs to be indexed and eligible for a snippet; eligibility does not guarantee selection. These requirements are explained in its official AI features guidance.

Before rewriting dozens of articles, I would look for obstacles affecting the website: important pages that are hard to reach, outdated information, confusing navigation or a mobile experience that makes reading difficult. The review should reflect the project's priorities, rather than apply an identical list of fixes to every business.

Our free SEO analysis tool provides a starting point for checking meta tags, structure and performance. Its report can highlight areas worth investigating. It does not certify that a page will be cited in an AI answer.

How to measure citations, clicks and business results

I would start with a documented baseline. For a relevant group of pages and searches, record clicks, impressions, click-through rate and website outcomes. Alongside that, observe AI visibility, recording the question, language, market, date and cited URL. One answer can be worth examining, but it is insufficient to describe a trend.

Click-through rate, or CTR, is clicks divided by impressions. If impressions increase while clicks stay unchanged, CTR can fall even though traffic has not declined. Equally, a higher CTR can accompany fewer clicks if demand or overall appearances fall. I would therefore read absolute numbers and percentages together.

According to Google's documentation, AI-feature appearances are included in overall Web search performance in Search Console. An organic increase alone does not identify clicks from one particular AI link. Assessing business value also requires tracking meaningful actions on the website.

Consider a hypothetical page moving from 1,000 visits and 10 enquiries to 800 visits and 16 enquiries. Traffic falls while enquiries increase. That deserves analysis, including the quality of those enquiries; it does not prove that AI caused the change. Focusing solely on visits would miss an important part of the picture.

Why a simple before-and-after comparison is not enough

If an article gets more clicks after October 15, timing alone does not establish that Google's new links deserve the credit. Rankings, demand, competitors, the page title or a brand-awareness campaign could have changed. Comparing just two weeks may capture ordinary fluctuation rather than a durable improvement.

I would compare equivalent periods, separating devices, countries, branded searches and generic queries. I would log changes made to the website and, where possible, observe comparable pages left unchanged. That does not create a perfect experiment, but it reduces the risk of attributing every movement to the latest announcement.

For lower-traffic websites, I would accept longer observation periods and more cautious conclusions. A handful of extra clicks can produce impressive percentages without yet representing a stable change.

Where GEO tracking fits into the strategy

Organic monitoring can be paired with visibility tracking across generative answers. Webita's SEO & GEO Tracking service brings together AI citations, keywords and supporting signals to help follow a project's development.

I would use that information to identify questions deserving attention and pages worth improving, keeping observed visibility separate from clicks and conversions. A visibility score is useful when it supports a decision; it cannot replace the website's business outcomes.

In my earlier article on GEO and AI Visibility, I explored monitoring in more detail. The next step here is connecting those observations to audience behaviour: who arrives, what they read and which actions they take.

Where I would start improving the content

I would not rewrite an entire blog in response to a change in Google's interface. I would select a manageable group of business-relevant pages, examine the questions they address, investigate lost clicks and review what they offer beyond the initial answer.

One page may need an updated comparison. Another may need a clearer explanation of costs, a documented process or an easier way to enquire. Internal links should support that journey: once readers understand the problem, they should be able to reach the relevant service or tool without friction.

That is the connection we work on through Webita's SEO consulting: aligning technical analysis, content and business goals rather than collecting pages around unrelated keywords. AI answers belong within that strategy alongside other forms of search.

The question I would ask goes beyond “Does Google cite me?” to “What useful information does someone find when they choose to open my website?”.

More prominent links may make that step easier. Publishers still need to provide a worthwhile reason to take it and a useful experience after the click. Comparing visibility, visits and outcomes is how we can assess what this evolution contributes to each individual project.