I saw it in my Google Search Console data in November 2025. Several of my highest-traffic informational pages — how-to articles about web design basics, explainer posts about SEO concepts — dropped 25–40% in clicks while maintaining similar or higher impressions. Same keyword ranking positions. Fewer clicks. The AI Overview was answering the question well enough that fewer people needed to visit the page.
I want to walk through exactly what changed, what I've tried, and what the data says about what's actually working now.
What my data showed (specific pages and changes)
I tracked 40 pages across my site for 6 months after AI Overviews expanded in India. The results were consistent with patterns reported across the industry, but not uniform:
Pages hit hardest: basic informational posts — "what is a meta description," "how does Google crawl a website," "what is a responsive website." These pages had CTRs drop from 4–8% to 1.5–3%. They still rank. Fewer people click through because the AI Overview answers the basic question in 3 sentences at the top of the page.
Pages largely unaffected: local service pages ("web designer in Pune Baner"), commercial content ("how much does a website cost in India"), and personal-experience posts ("I built a client website in an hour — here's how"). These pages either don't trigger AI Overviews (local queries rarely do) or their content is inherently more specific than what an AI summary can provide.
Pages that actually gained traffic: detailed comparison posts with my specific experience ("Cursor vs Copilot — what I actually use"), posts with original data or tested outcomes, and posts that rank for queries where the AI Overview links out to my site as a source.
The content strategy shift that's actually working
I've effectively stopped publishing "what is X" and "how to do basic Y" content. Not because it's wrong to publish — but because the return on that kind of content is collapsing. The AI does it adequately for the majority of searches. Publishing generic informational content in 2026 to capture search traffic is like opening a payphone business in 2010. The market is there today, but the trajectory is clear.
The content type I'm investing in instead: documented experience and original perspective. "I tested X, here's what I found." "I've done this 40 times, here's the pattern." "Here's what the official guides don't tell you that I learned from working with 15 clients on this." This content has something AI Overviews can't synthesise: actual first-person specificity that adds value beyond the general answer.
The SEO lens: queries where the searcher wants to know what someone who has actually done the thing thinks — not a summary of publicly available information — still drive clicks to the people who can provide that. Establish yourself as that person in your niche.
Technical adaptations that helped
Adding FAQ schema to all key pages. When your FAQ content is structured and marked up correctly, it can appear as part of or adjacent to the AI Overview — sometimes with a link. This has partially restored visibility on some pages that were losing raw clicks.
Building more internal linking between informational content and commercial pages. The informational content may get fewer clicks, but if it still drives some traffic and links to a service page or a conversion-focused post — the informational content still has value in the funnel even if it no longer terminates the journey on its own page.
Prioritising featured snippet optimisation for the informational pages I kept. If you're going to maintain an informational page, at minimum aim for it to be the source Google's AI Overview cites — which requires clear, structured, authoritative content and appropriate schema.
The longer-term view
For local service businesses in India — the kind I build websites for — AI Overviews are a smaller concern than for content publishers. Local queries ("plumber in Baner," "CA Andheri," "interior design Koramangala") still show traditional organic results and Google Maps results. The people searching those queries need a local business to call — an AI summary doesn't answer that need. If your website is primarily driving local enquiries rather than informational traffic, the AI Overview impact is manageable.
For content-dependent businesses — affiliate sites, blogs monetised by traffic, businesses whose lead gen depends on informational content — the adaptation is non-optional. The informational content layer is being abstracted away by Google's AI. The next layer — genuine expertise, documented experience, original perspective — is where the traffic and credibility go next.
How to monitor AI Overview impact in Search Console
The data you need is in Google Search Console, but the default view obscures it. Here's how I actually track AI Overview impact:
Filter Search Console performance data to the past 16 months (the maximum range) and look at impressions vs clicks over time, not just for your site as a whole but per individual page. The signature of AI Overview impact is: impressions holding steady or growing (your ranking position is maintained) while clicks drop significantly (fewer people need to visit). A page where impressions increased 20% but clicks dropped 30% is almost definitely experiencing AI Overview interference.
Cross-reference this with the keywords. Go to the Queries tab, filter by the pages you suspect are affected, and look at which queries show high impressions and near-zero CTR. These are the queries where an AI Overview is answering the question well enough that click-through is rare. That's your inventory of "commoditised keywords" — informational queries your content no longer effectively owns as traffic sources.
The actionable step: for each page showing this pattern, decide whether to (a) rewrite the page to focus on comparative or experiential angles that the AI Overview doesn't satisfy, (b) redirect the keyword strategy to related commercial-intent queries on the same topic, or (c) accept the reduced traffic and focus on internal linking to conversion pages from whatever residual traffic arrives.
The pages that actually recovered — what they had in common
I've spent some time studying which pages in my network recovered impressions and CTR after initial AI Overview impact. The pattern is pretty clear:
Pages with first-person specificity recovered. A post about "how to budget for a website in India" that included specific client quotes, real ₹ numbers from actual projects, and my personal opinion on what's overpriced — that page dropped initially and then recovered. The AI Overview that appeared for that query is generic. Users who see the generic summary and want the specific Indian practitioner perspective still click through.
Pages with original data or structured comparison tables recovered. A post with a table I built from personal testing — comparing 5 web builders on 8 India-specific criteria — held its CTR because the AI can't replicate my tested outcomes. It might cite my table (sometimes does), but the full structured reference is still on my page.
Pages targeting niche or hyper-specific queries held very well. "Web designer in Baner" or "website cost for restaurant Hyderabad" barely changed. AI Overviews appear infrequently for specific local or niche queries because there isn't enough consistent signal from user queries to train the overview behaviour.
What didn't recover: the truly generic informational posts. "What is bounce rate" and "what does meta description do" — these are dead as primary traffic sources. I've accepted that and redirected my content energy to topics where my lived experience adds something the AI layer cannot synthesise.
Also see: Does Google rank AI-written content in 2026? and How to automate your SEO and lead generation.