How to Combat Traffic Loss from Google's AI Overviews

What changed in Google's search results?

Google now answers some queries directly with AI Overviews, placing a generated response and supporting links above or alongside familiar organic results. The change matters because a searcher can get a usable answer before visiting a publisher or brand site. Google says AI Overviews appear only when its systems judge them additive to classic Search, and that the has can show links to supporting pages. Google's guidance for site owners also says that AI Overviews and AI Mode may use query fan-out, which runs related searches across subtopics and data sources.

The practical change is not that organic ranking disappeared. It is that ranking, impression volume, click-through rate, and business value can now move independently. A page may remain visible in Search Console while a larger share of searchers stop at the results page. That makes a traffic-only dashboard incomplete for teams whose category questions now trigger AI Overviews.

The current debate also has a second force behind it: Google continues to enforce its spam policies while generative tools make generic search copy easier to produce. The August 2026 report from Storyboard18 describes an August spam update that rolled out from August 18 through August 21, alongside growing concern about AI Overview click loss. Treat those as related market conditions, not as proof that every traffic decline came from one algorithm update.

The important distinction is causal discipline. An AI Overview appearing on a query does not prove it caused every lost session. Seasonality, ranking changes, indexing problems, SERP features, demand shifts, technical releases, and spam-policy enforcement can all affect traffic. Teams that label every decline as an AI Overview problem will make the wrong fix with great confidence.

1. Map the query set where AI Overviews appear before changing content strategy.
2. Separate lost clicks from lost rankings, lost impressions, and lost conversions.
3. Treat broad spam enforcement as a quality risk, not as a license to publish more generic pages.

Who is most exposed to AI Overviews traffic loss?

Sites that depend on informational queries are most exposed because AI Overviews are designed to help people get the gist of a complex topic or question quickly. Ahrefs found that 99.2% of keywords triggering AI Overviews in its sample were informational, which is why its click-through analysis focused on informational terms. Its methodology compared 150,000 keywords with an AI Overview against 150,000 informational keywords without one.

That puts publishers, B2B SaaS content teams, affiliate sites, local service businesses, and comparison-led brands under pressure. The common risk is a page that answers an early research question but has little reason to continue. If the searcher only needed a definition, a short list, or a basic comparison, the AI Overview can satisfy the need before the visit.

Exposure is not the same as inevitability. A local business can still earn a visit when a searcher needs availability, location context, pricing, proof, or a booking path. A B2B buyer still needs product detail, implementation evidence, security information, and a reason to trust a vendor. The challenge is that thin pages built to capture the first click are less defensible when the results page handles the first answer.

The teams at greatest risk are those that report only aggregate organic traffic. Google includes traffic from AI has in the Web search type in Search Console, rather than exposing a separate AI Overview traffic bucket. That means a dashboard can show a decline without revealing which queries, pages, or result layouts changed. Google recommends using Search Console alongside Analytics and tracking conversions and time spent on site. The documentation says clicks from results pages with AI Overviews may be higher quality, but each site needs to test that against its own conversion data.

1. Informational pages with weak next steps face the largest click risk.
2. Pages serving high-intent research can still win visits when they provide detail the overview cannot finish.
3. Aggregate reporting hides the query-level evidence needed to respond.

What does the before-and-after evidence show?

The before-and-after evidence shows why a stable top ranking can still produce less traffic. Ahrefs compared March 2024, before the US rollout of AI Overviews, with March 2025 for a 300,000-keyword dataset. For keywords that triggered an AI Overview in March 2025, the average position-one click-through rate fell from 7.3% to 2.6%. After adjusting against the broader decline in informational click-through rate, Ahrefs estimated a 34.5% reduction in position-one click-through rate associated with AI Overviews. Read the study and methodology.

That result is evidence of correlation within one study, not a universal traffic forecast. It does not mean every number one page loses 34.5% of clicks, and it does not isolate every change in search behavior. It does make one point hard to ignore: a page can retain prominent organic placement while its expected click yield falls.

The appropriate response is to use the same before-and-after logic on your own data. Select a stable set of pages and queries, record the period before visible AI Overview exposure, then compare it with a later matched period. Keep rankings, impressions, clicks, click-through rate, conversions, and assisted conversions separate. Add an observation of whether the query currently shows an AI Overview, because its absence is as informative as its presence.

Do not use a traffic decline to justify untested content production. Start with pages that have high impressions, material click loss, and a clear commercial or audience role. Those pages give you a practical place to improve the answer, strengthen evidence, and create a path beyond the first search interaction.

1. Use a matched before-and-after period for the same query set.
2. Check ranking stability before assigning lost clicks to AI Overviews.
3. Prioritize pages by lost business value, not by traffic loss alone.

How should you diagnose an AI Overviews traffic decline?

Diagnose AI Overviews traffic by building a query-level baseline before rewriting pages. Export Search Console Web performance data for a meaningful pre-change and post-change period. Group queries by intent, page, device, country, and brand status where the volume supports it. Then inspect the live results for a representative set of declining queries and record whether an AI Overview appears, which sources it cites, and what the response actually answers.

A clean diagnostic separates four outcomes. First, impressions may fall because demand or eligibility changed. Second, rankings may fall because competitors or Google systems changed. Third, click-through rate may fall while rankings remain stable, which is consistent with a changed results page. Fourth, clicks may decline while conversion rate rises, suggesting fewer but more qualified visitors. These outcomes require different actions, so they should not share one generic remediation plan.

Google's advice does not call for a new technical file, special AI markup, or a separate schema type for AI Overviews. A page needs to be indexed and eligible to show a snippet in Google Search, and existing technical requirements and SEO best practices still apply. Google explicitly says there are no additional technical requirements for eligibility as a supporting link.

This is also where many teams make an expensive mistake. They create pages that mimic the wording of an AI Overview, then publish many near-identical variants. That may add little new information and increases quality risk. A stronger audit asks what the page can contribute that a generated summary cannot: primary research, a dated methodology, first-party product evidence, expert review, local detail, or a transparent comparison.

1. Export Search Console data before editing the affected pages.
2. Inspect live SERPs manually and preserve a dated record of cited sources.
3. Measure conversions and assisted conversions alongside clicks.
4. Fix indexing, snippet eligibility, and content quality before seeking a new tactic.

How should content change after an AI Overview appears?

Content should become easier to cite and harder to replace. Start with an answer-first opening that states the direct conclusion, then show the evidence, scope, date, and limitations that make the answer trustworthy. Use clear headings that match real questions. Add a comparison table when the reader must choose between options. Publish original data only when the collection method can be explained and the numbers can be checked.

This is the operating logic behind Citation Engineering: create authoritative coverage that AI systems can understand and support with a citation, rather than attempting to manipulate a model. The mechanism is quality, coverage, and freshness. A page that merely restates common knowledge may be useful, but it has a weaker claim on citation than a page with a named source, a current date, a clear method, and a distinct point of view.

Use the overview itself as research, not as a template. Note the question, the sources shown, the missing details, and the follow-up question a serious buyer or customer would ask next. Build a page that answers that follow-up with evidence. For a service business, that may mean a location-specific explanation and a route to contact. For B2B SaaS, it may mean implementation constraints, comparison criteria, or a proof-backed use case.

Keep the user experience intact. If an AI Overview summary earns the first answer, the destination page must earn the second step. Make the commercial next action clear, but do not turn every informational page into an interruption. A strong page lets a reader verify a claim, understand what changes their decision, and continue when they are ready.

1. Open with the direct answer and state the evidence boundary.
2. Add dated sources, methods, tables, or original data where they help the reader decide.
3. Build pages around the next question, not a rewritten version of the overview.
4. Connect informational content to a relevant owned conversion path.

Which metrics should replace a traffic-only scoreboard?

Traffic should remain on the scoreboard, but it should no longer stand alone. Track Citation Share, the percentage of relevant AI answers in a category that cite your brand. Pair it with Answer Presence, which shows breadth across the question universe, and Citation Count per day, which shows volume. These measures show whether a brand appears in the answers that shape consideration, not only whether it earns a traditional blue-link click.

Add Search Console clicks, click-through rate, conversions, assisted conversions, and engagement quality to the same reporting view. Google says AI has traffic is reported in the Performance report under the Web search type, while Google Analytics can help assess conversions and time on site. That creates a practical measurement stack: Search Console identifies search behavior, Analytics identifies site outcomes, and citation tracking identifies whether the brand is present in AI answers.

A final metric deserves attention: Share of Voice against named competitors. If a category question repeatedly cites rival brands and excludes yours, a healthy aggregate traffic chart will not solve the strategic problem. The aim is not a vanity mention. It is sustained inclusion where a buyer asks a relevant question and an answer engine chooses sources.

The shift is uncomfortable because it removes the simplicity of one rank and one traffic number. It also creates a clearer job for editorial teams. Publish material worth citing, observe where it earns inclusion, and connect that visibility to measurable business outcomes. Rankings got you found. Citations help determine whether you are chosen.

1. Keep organic traffic, but report it beside conversions and engagement quality.
2. Track Citation Share for category and comparison prompts.
3. Use Answer Presence to find coverage gaps across the question universe.
4. Compare citation visibility with named competitors before declaring a content program healthy.

Sources

Source: Google Search Central

Google's site-owner guidance states that AI Overviews and AI Mode use Search eligibility and existing SEO best practices, with no additional technical requirements for supporting-link eligibility. Google Search Central, 2025-12-10

Source: Ahrefs

Ahrefs analyzed 300,000 keywords and estimated that AI Overviews reduced position-one click-through rate by 34.5%, comparing March 2024 with March 2025. Ahrefs, 2025-04-17

Source: Storyboard18

Storyboard18 reported that Google's August 2026 spam update rolled out from August 18 through August 21 amid concern about declining publisher traffic and AI Overviews. Storyboard18, 2026-08-27