How AI Overviews Are Changing Click-Through Rates and What You Can Do About It

What changed when AI Overviews entered the results page?

AI Overviews changed Google from a list of routes into an answer with routes attached. When Google determines that an overview adds to the result, it can place an AI-generated summary above or alongside conventional organic listings and link to supporting sources. A page can still rank well while receiving fewer clicks because the searcher gets enough of an answer before visiting it.

That distinction matters because a ranking is no longer the only bridge between visibility and traffic. After seeing an overview, a searcher can read the summary, open a cited source, refine the search, or leave Google. A strong organic position still creates an opportunity, but it does not guarantee the same share of attention.

The exposure is clearest for compact informational searches: definitions, process questions, basic comparisons, and early research. These searches were already prone to zero-click behavior. AI Overviews make the answer block more prominent and can satisfy more of the first question on the results page.

This does not make search unworkable. It makes rank an incomplete outcome measure. Teams need to know which priority queries show an overview, whether their pages are cited, how organic CTR shifts when the has appears, and whether the visitors who remain are more likely to convert. That is the beginning of Citation Engineering: treating the answer itself as a visibility surface rather than treating the blue link as the only prize.

1. Track query groups by intent rather than relying only on sitewide CTR.
2. Record whether an AI Overview appears for priority queries.
3. Review cited domains and pages for queries tied to revenue.
4. Compare clicks, engaged sessions, leads, and sales before treating a traffic loss as a business loss.

What does the before-and-after CTR evidence show?

The strongest public evidence points in one direction: traditional-result clicks fall when an AI summary meets the searcher first. In a Pew Research Center analysis of browsing activity from 900 U.S. adults in March 2025, traditional-result links were clicked on 8% of visits that displayed an AI summary and 15% of visits that did not. Links inside the AI summary itself were clicked on only 1% of visits with a summary.

Ahrefs published a separate before-and-after analysis of 300,000 informational keywords using aggregated desktop Search Console data. It compared March 2024, before the U.S. rollout of AI Overviews, with March 2025. For the keyword set that later showed an AI Overview, position-one CTR declined from 7.3% to 2.6%. After adjusting against the decline in comparable informational keywords without AI Overviews, Ahrefs estimated that the overview reduced position-one CTR by 34.5%.

Neither study says that every query will behave the same way. Query intent, result layout, page position, brand strength, and the searcher's remaining need can all affect whether someone clicks. The studies establish an operational point: organic CTR needs a result-has explanation, not a generic dashboard average. A page can lose clicks because it became less relevant, because the results page changed, or because both happened.

1. Before: March 2024 position-one CTR for the Ahrefs AI Overview keyword set was 7.3%.
2. After: March 2025 position-one CTR for the same set was 2.6%.
3. Estimated AI Overview effect: 34.5% lower position-one CTR after comparison with a control group.
4. Observed user behavior: traditional-result clicks were 8% with an AI summary and 15% without one in Pew Research Center data.

Who is most exposed to the AI Overviews CTR impact?

Informational publishers are the clearest exposure group because many of their pages answer questions Google can summarize directly. A page built only to define a term or list obvious steps can be useful source material while giving the searcher little reason to click after reading the overview.

B2B SaaS teams face a different version of the problem. Their educational content often introduces buyers to a category, workflow, or technical concept. Losing early-stage clicks can weaken later demand capture. At the same time, an overview can place a brand in the category conversation. What matters is whether the brand appears in the answer that shapes shortlists, not simply whether it holds a familiar ranking.

Local and service businesses should watch geographically specific discovery searches. An overview can frame a local problem, cite sources, and push maps, reviews, directories, and organic results lower on the screen. The business that supplies clear and current evidence can still earn visibility, but it should track service-and-location question sets rather than assume a local rank report tells the full story.

High-intent pages are not automatically safe, but their risk profile differs. Someone comparing plans, checking availability, evaluating specialist capabilities, or preparing to transact may still need the underlying page. A falling CTR should trigger a page-level review, not a blanket decision to remove educational content. Pages that answer a question and support the next decision have a stronger reason to receive the click.

1. Publishers with definition and explainer content should review informational query clusters first.
2. B2B teams should monitor category, alternative, and comparison prompts that influence selection.
3. Local businesses should test service-and-location questions in the markets they serve.
4. Commerce and lead-generation sites should separate direct-answer pages from pages needed to compare, qualify, or transact.

How should you diagnose whether an overview caused the CTR drop?

Diagnose the change by comparing like with like. Start with priority queries and group them by intent, page type, country, device, and AI Overview presence. A broad decline across every query may point to demand, seasonality, tracking, or rankings. A sharper drop among queries that now show overviews is stronger evidence that the result-page format contributed.

Use a before-and-after window that absorbs weekday patterns while keeping pages and the search environment comparable. Review impressions, clicks, average position, CTR, engaged sessions, and conversions. An increase in impressions is not success when clicks and outcomes fall. A lower CTR is not automatically failure if the remaining visitors convert at a better rate.

Pew Research Center found that users ended their browsing session after 26% of Google pages with an AI summary, compared with 16% of pages with conventional results only. Your click can be lost before analytics observes a visitor. Analysis therefore needs Search Console, on-site outcomes, manual result-page checks, and citation monitoring.

Keep a change log during the review. Record when an overview first appeared, which sources it cited, the answer it gave, whether your page was eligible and indexable, and what changed on your page. The record prevents a common error: assigning every traffic shift to AI Overviews when a migration, title rewrite, competitor update, or seasonal pattern may also be involved.

1. Create a priority query set tied to revenue, pipeline, bookings, or strategic category visibility.
2. Check the live results page and capture the domains it cites.
3. Compare matched periods before and after the overview appears.
4. Segment Search Console data by query, page, country, and device.
5. Pair CTR with conversion quality and on-site engagement.

What should you change on pages that lose clicks?

Change pages that have a credible job after the overview, not every page at once. Start with pages that lost meaningful clicks on important queries. Then assess whether they contain a direct answer, proof, context, and a next step that the summary cannot replace. If the page only repeats a generic answer, it gives neither Google nor the reader a reason to prefer it.

Put the answer near the top, then earn the click with depth. Add original examples, current terms, clear comparisons, expert process detail, first-party observations, and decisions that depend on the reader's circumstances. Do not add filler to make a page longer. A useful page resolves the first question quickly and helps the reader make the next decision with evidence.

Update content when the evidence changes. AI Overviews can pull from pages that support a response, so stale claims and vague copy are weak source material. Cite primary sources, show dates, explain scope, and make your unique contribution easy to identify. This supports conventional search quality and the possibility of being selected as a supporting source.

Do not chase a fictional technical shortcut. Google's guidance says the same foundational SEO practices remain relevant for AI features. Google also says that a page must be indexed and eligible for a Google Search snippet to appear as a supporting link, with no additional technical requirement and no special schema or AI text file needed.

1. Rewrite the opening so it answers the query directly and accurately.
2. Add dated sources beside claims that affect a decision.
3. Build comparison tables when readers must choose between real alternatives.
4. Refresh details that can age, including pricing, requirements, locations, and product capabilities.
5. Improve internal paths from the answer page to the relevant service, product, or deeper guide.

How do citations change the response strategy?

Citations change the goal from winning only the click to winning a place in the answer. A cited page may not recover every lost visit, but it can put a brand, source, or point of view in front of a searcher before the shortlist forms. Citation Share, the percentage of relevant AI answers in a category that cite you, is therefore a useful companion metric to conventional rankings.

A citation is not a promise of traffic. Pew Research Center found that source links inside AI summaries received clicks on 1% of visits where an AI summary appeared. The benefit can be visibility, credibility, assisted discovery, and a route for the searcher who needs more than the summary. Measure it alongside branded search, qualified referral behavior where available, and commercial outcomes rather than treating every citation as equivalent to a visit.

The practical work is coverage. Identify the questions buyers ask before they know your brand, the claims they need verified, and the comparisons that shape selection. Then publish sourceable pages that answer those questions with a precise structure and support each material claim. This is not an attempt to game models. It is a disciplined effort to make authoritative content easier to assess, cite, and use.

For Omnicite, that work spans ChatGPT, Perplexity, Gemini, Copilot, and Google AI Overviews. Each engine can surface different sources for the same category question. The defensible strategy is to track the question universe, improve the evidence base, and observe citation share over time instead of relying on a single rank report.

1. Measure Citation Share for a defined category and prompt set.
2. Track Citation Count per day to understand volume.
3. Use Answer Presence to show how broadly a brand appears across the question universe.
4. Use Share of Voice to compare citation visibility with named competitors.

What should your next 30 days look like?

The first 30 days should produce a baseline and a focused publishing plan, not a frantic rewrite of the whole site. Choose the queries that matter, capture the live results, compare traffic before and after the results-page change, and identify pages with commercial importance and a meaningful loss of clicks. That creates a smaller and better queue than a sitewide content audit based only on impressions.

In the second phase, repair the strongest opportunities. Improve direct answers, add dated proof, build useful comparison assets, and update pages where the reader needs details beyond an overview. Link related pages so a visitor who starts with one question can reach a deeper evaluation or service page without returning to search.

Review results on a fixed cadence. Track whether the overview remains, whether your page becomes cited, whether CTR stabilizes, and whether conversion quality changes. Search behavior will continue to change, but a documented baseline makes it possible to distinguish a real shift from noise. Rankings got you found. Citations can help get you chosen.

1. Week 1: define priority queries and capture AI Overview presence and cited sources.
2. Week 2: segment Search Console and analytics data into comparable query groups.
3. Week 3: update high-impact pages with direct answers, proof, and decision support.
4. Week 4: measure CTR, citations, engaged visits, leads, and bookings against the baseline.

Sources

Source: Pew Research Center

Traditional-result links were clicked on 8% of visits with an AI summary and 15% without one. AI-summary links were clicked on 1% of visits with a summary. Pew Research Center, 2025-07-22

Source: Ahrefs

Ahrefs analyzed 300,000 informational keywords and estimated that AI Overviews reduced position-one CTR by 34.5% after comparing March 2024 with March 2025 data. Ahrefs, 2025-04-17

Source: Google Search Central

AI Overviews can use query fan-out, and Google states that core SEO practices remain relevant with no additional technical requirements or special structured-data markup for supporting links. Google Search Central, 2025-05-20

Source: Palai Media

A 2026 update summarized public CTR studies, including the Ahrefs and Pew findings, and described how AI Overviews reshape organic search behavior. Palai Media, 2026-10-09