Do AI Overviews Reduce Clicks? What to Measure

What AI Overviews are

• Definition: AI Overviews are AI-written summaries that Google shows at the top of some search results, with links to sources.
• Related feature: AI Mode is a separate, more conversational Google experience.
• How they are chosen: Google says its AI features rely on its normal Search systems. Pages must be indexed and eligible to appear with a snippet.
• Not on every search: They appear more often on informational queries than on local, branded or transactional ones.
• They change: Google updates where and how often they appear, so any study is a snapshot.

What the studies found

Image slot 2 · 1200x800 pxai-overviews-ctr-studies.webp

Alt text: Two bar charts comparing click-through rate with and without AI Overviews from two separate studies

AI image prompt

"Two glass chart cards side by side on a dark navy background, each with a pair of vertical bars where one is clearly shorter, in cyan and blue with a green highlight. Subtle grid, soft glow, minimal, no readable text, no logos."

Copy prompt

Why the numbers differ

• Different measures: Ahrefs forecasts what position 1 CTR would have been without AI Overviews. Seer compares groups of queries over time.
• Different data: Ahrefs uses a broad keyword set. Seer uses 42 client accounts, so industries and brands skew the result.
• Different periods: The studies cover different months, and AI Overviews changed over that time.
• Device: The Ahrefs study is desktop only.
• Baseline: Both show that CTR also fell on queries without AI Overviews, so some decline is not caused by them.
• Method choices: Ahrefs's forecast assumes keywords without AI Overviews are a fair stand-in for what AI Overview keywords would otherwise have done. The study does not test that.
• Takeaway: Treat both as a range of evidence, not one answer.

What the studies cannot tell you

• They are observational. They show association, not proof that AI Overviews caused each drop.
• They are vendor-run. Ahrefs and Seer sell SEO tools and services. That does not make them wrong, but you should read the limits.
• They cover informational queries. Seer states results may differ for commercial, transactional or local searches.
• They show averages. Seer reports wide variation between queries and industries.
• They don't measure business value. A lost click is not necessarily a lost customer, and a click is not necessarily a sale.
• Some figures conflict. Seer's report contains small internal inconsistencies, so we used its main data table and rounded figures.
• They go out of date. Both describe past behaviour of a changing feature.

Does being cited help?

• What Seer found: In Q3 2025, organic CTR was 0.70% when a brand was cited in the AI Overview and 0.52% when it wasn't. That is about 35% more clicks.
• Paid too: The same pattern held for paid clicks, though the paid sample was smaller.
• The caveat: Seer says it cannot prove citation causes higher CTR. Stronger brands may be both cited more and clicked more.
• Practical reading: Being cited is worth pursuing. It is not a substitute for ranking well.
• Related guide: See how to get cited by ChatGPT, Gemini and Perplexity.

Which searches are affected most

• Informational questions: "What is…", "how to…", "why does…" are most likely to show an AI Overview.
• Top positions: Ahrefs reported the largest drop at position 1 (58%), shrinking at lower positions: about 51% at position 2 and 19% at position 10.
• Less affected: Local searches ("near me"), branded searches and many transactional searches, though this varies and changes.
• Your own mix matters: If most of your traffic comes from how-to content, you're more exposed than a business that earns most leads from local and service queries.

What to measure on your own site

Image slot 3 · 1200x800 pxai-overviews-measurement-dashboard.webp

Alt text: A simple dashboard showing impressions, clicks, click-through rate and enquiries over time

AI image prompt

"A clean glass dashboard card on a dark navy background with four small metric tiles above a two-line chart, in cyan, blue and green. Subtle glow, rounded, minimal, no readable text, no logos."

Copy prompt

• Impressions: Are people still seeing your pages?
• Clicks and CTR by page group: Informational pages versus service and local pages.
• Average position: To separate ranking drops from click drops.
• Branded searches: A stable or rising brand search count suggests awareness is holding.
• Enquiries and calls: Form submissions, calls, WhatsApp clicks and direction requests.
• Conversion rate: If clicks fall but enquiries hold, the lost clicks may have been low value.
• AI referral visits: Sessions from AI platforms, noting some are not labelled.
• AI citations: Monthly prompt tests for your key questions.

How to measure it step by step

• Step 1: Group your pages. Split into informational (guides, blog) and commercial (services, local pages).
• Step 2: Set a baseline. Export 12 months of Search Console data by page group: impressions, clicks, CTR, position.
• Step 3: Compare year on year. Compare the same months to avoid seasonality.
• Step 4: Find the pattern. Impressions steady but CTR down on informational pages is consistent with answers appearing on the results page. Position falling points to a ranking issue instead.
• Step 5: Check query examples. For your top queries, search them yourself and see whether an AI Overview appears. This varies by location and time, so repeat on different days.
• Step 6: Connect to enquiries. Match changes against leads from your CRM or inbox.
• Step 7: Test AI answers. Run your question list in AI tools monthly and log citations.
• Step 8: Review quarterly. Make decisions on trends over months, not days.

What Search Console can and cannot show

• It can show: Impressions, clicks, CTR and position for your pages and queries.
• Generative AI report: Google has added reporting for its AI features. As far as we can tell, it does not break out query-level data and combines AI Overviews and AI Mode. Check Google's current documentation, as this may change.
• It cannot show: A clean "AI Overview clicks" number for your site, or exactly which searches displayed one to each user.
• Third-party tools: Rank trackers can flag AI Overview presence for keyword sets. Treat them as samples, because results vary by location and user.
• Be transparent: In reports, say which numbers are measured and which are estimated.

What to do about it

• Protect high-value pages: Prioritise service, local and comparison pages that lead to enquiries.
• Answer first, then add value: Give the short answer, then offer what an AI summary can't: your examples, pricing context, photos, checklists and tools.
• Aim to be cited: Clear passages, named sources and consistent facts help. See the AEO checklist.
• Build brand demand: If people search for you by name, you're less dependent on generic clicks.
• Strengthen your Google Business Profile: Local discovery often happens there. See Google Business Profile.
• Capture value without a click: Make your name, phone number and key facts show correctly in the results.
• Diversify: Email, WhatsApp, referrals and partnerships reduce dependence on one source.
• Don't panic-delete content: Informational pages build authority and feed citations.

Small and local businesses in Mumbai and India

• Local intent is less exposed: Searches like "CCTV installation Andheri" usually show map packs and listings, so Google Business Profile matters.
• Informational content still matters: It builds trust and gets cited, even if it earns fewer clicks.
• Measure enquiries, not just visits: For a service business, 20 good enquiries beat 2,000 casual visits.
• Watch mobile: Most Indian search traffic is mobile, and AI features can take up much of a phone screen.
• No India-specific studies cited here: The studies we reviewed are global or US-weighted. We haven't found reliable India-specific CTR data.

Mistakes when reading the data

• Treating "58%" as your number: It's a study average for one condition, not a forecast for your site.
• Ignoring non-AI decline: CTR also fell where there was no AI Overview.
• Comparing studies directly: They use different methods and periods.
• Judging from a single query: AI Overviews vary by user, place and day.
• Only counting clicks: Visibility and enquiries matter too.
• Blaming AI for everything: Rankings, seasonality, site changes and competitors also move traffic.

Limits and honest caveats

• We haven't run our own study. We summarise published research and explain its limits.
• Studies change. Vendors update their data, and the feature itself changes.
• Google's own view differs. Google has publicly said that AI features can send quality clicks, but we haven't found independent data that settles the question, so we don't quote it as fact.
• No guarantee. We can't promise any change will recover clicks or earn citations.
• Bias note. We sell AI search optimization services.

How our services help

• SEO services: Page-group reporting that separates ranking changes from AI-related click changes.
• AI search optimization: Citation testing and answer-ready page improvements.
• Google Business Profile management: Local visibility that doesn't depend on blue-link clicks.
• Website development: Tracking setup so enquiries are measured properly.