If your site's traffic dropped this year, AI Overviews probably ate it. Panicking won't bring those clicks back, but shifting from ranking pages to Answer Engine Optimization will.
Something broke in analytics dashboards across the internet this year, and it was not a tracking bug. Organic traffic for informational queries dropped off a cliff. Not a gentle decline. Not the usual algorithm shuffle that SEOs have learned to ride out every few months. A structural change.
The culprit is Google's AI Overviews. Those AI-generated summary blocks that now appear at the top of search results for a growing number of queries. A Pew Research Center study found that users click traditional search result links just 8% of the time when an AI Overview appears, compared to 15% when there is no overview. Users click links inside the AI Overview itself only 1% of the time. If your content strategy was built around ranking for "what is X" queries, those clicks vaporized into a summary box that answers the question without anyone needing to visit your site.
The instinct is to panic. The productive response is to understand what actually changed and rebuild your strategy around the new reality.
Zero-click searches are not new. Featured snippets have been eating clicks for years. But AI Overviews are a different animal. Featured snippets pulled a sentence or two from a single source, and users often clicked through for the full answer. AI Overviews synthesize information from multiple sources into a comprehensive summary. The answer is complete. The user has no reason to keep going.
And it is getting worse. According to Semrush Sensor data, AI Overviews currently appear for around 13% of search queries in the US. That number has been climbing steadily. More concerning: the types of queries triggering AI Overviews are expanding fast. In October 2024, 89% of keywords triggering overviews were informational. By October 2025, that dropped to 57%. Commercial and transactional queries are increasingly getting AI-generated summaries too.
| Query type | AI Overview frequency (Oct 2024) | AI Overview frequency (Oct 2025) | Direction |
|---|---|---|---|
| Informational | 89% of triggered queries | 57% of triggered queries | Declining share (but not declining volume) |
| Commercial | ~6% | ~22% | Growing fast |
| Transactional | ~3% | ~13% | Growing fast |
| Navigational |
That table tells the real story. Google is not just summarizing encyclopedia-style questions anymore. It is summarizing product comparisons, buying guides, and "best X for Y" queries. If your traffic came from any of those categories, the ground has shifted under you.
Here is a pattern that used to be fine: you publish five articles about slightly different angles of the same topic. "What is Kubernetes?" "Kubernetes for beginners." "How Kubernetes works." "Kubernetes vs Docker." "Why use Kubernetes in 2026." Each one targets a marginally different keyword. Each one ranks for its own slice of search traffic.
In the pre-overview world, this worked. Google would show the most relevant article for each query. You captured traffic across all of them.
In the AI Overview world, this backfires. Google's generative engine looks at your five articles and cannot figure out which one is your definitive source on the topic. It might pull sentences from two of them, ignore the other three, and cite a competitor who published one thorough pillar page that covered everything. Your authority got diluted across five URLs instead of concentrated in one.
Content consolidation has gone from "nice to have" SEO hygiene to a survival strategy. The sites recovering fastest from AI Overview traffic loss are the ones that merged their overlapping content into single, authoritative pages. One URL, one topic, maximum depth.
Before consolidating articles, audit which URLs have earned backlinks. Redirect the lower-authority pages to your consolidated pillar page using 301 redirects, so the link equity transfers. Deleting pages without redirecting destroys the signals Google uses to determine authority.
Pull a list of all pages ranking for similar keyword clusters. Look for pages competing against each other in Search Console (multiple URLs appearing for the same query). Any cluster with three or more overlapping pages is a consolidation candidate.
Pick the URL with the most backlinks, highest authority, or best existing rank. This becomes your single source of truth for that topic. Everything else redirects here.
Do not just redirect and delete the extras. Pull the best sections, data points, and unique angles from each overlapping article into the pillar page. The consolidated page should be more comprehensive than any individual piece was on its own.
Generative search engines parse content differently than humans. A human reader is fine with a 300-word preamble before getting to the actual answer. An AI Overview generator is not. It scans for information density and direct answers. If your article buries the answer in paragraph six after a long anecdote, the AI will pull content from a competitor who put the answer in the first sentence of the section.
This does not mean you write like a robot. It means you front-load the answer and then provide context. The structure that works best in 2026 follows a pattern: lead each H2 section with a 40-to-60-word direct answer to the implied question, then expand with analysis, examples, and nuance below it.
The schema markup game has also changed. FAQ schema, Article schema, and Dataset schema tell Google's generative engine exactly what your content contains and how it is structured. Sites that implemented comprehensive schema markup before AI Overviews rolled out wide have seen significantly better citation rates in AI-generated summaries.
| Schema type | When to use | Impact on AI Overviews |
|---|---|---|
| FAQ (FAQPage) | Pages answering multiple related questions | High. Direct Q&A pairs are easy for AI to parse and cite |
| Article | Blog posts, guides, news articles | Medium. Helps classify content type and publication authority |
| HowTo |
This is also where retrieval-augmented generation patterns become relevant to SEO. Google's AI Overviews use a process the company calls "query fan-out," running multiple related searches and combining the results. Your content needs to be the kind of well-structured, high-density source that a retrieval system would rank as authoritative for a given sub-query. Thin content with weak structure gets ignored even if it technically ranks on page one.
This is the hard truth that a lot of content teams do not want to hear. If your entire editorial calendar consists of "explainer" content that restates publicly available information in slightly different words, AI Overviews have made that content category nearly worthless for traffic acquisition. The AI can write that explainer itself. It does not need your version.
The content that still drives clicks through AI Overviews has a specific profile: it offers something the AI model could not generate from its training data. That means:
Original benchmarks and test results. If you ran a performance test comparing five CI/CD platforms and published the numbers, that data did not exist until you created it. AI cannot fake it. When the overview references CI/CD performance, your benchmark is the only primary source.
First-party case studies with specific metrics. "We migrated from MongoDB to Postgres and here is what happened to our query latency" is content that only you can write. It carries the weight of real experience, and Google's E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) rewards it accordingly.
Proprietary research and surveys. If you surveyed 500 developers about their AI tool preferences and published the results, that is a citable dataset. AI Overviews pull from primary research because the generative model cannot invent survey results.
Expert analysis with named attribution. Not "experts say" but "Jane Chen, principal engineer at Stripe, argues that..." Named sources with verifiable credentials make your content harder for AI to replicate and easier for Google to trust.
Brand mentions are becoming more important than traditional backlinks for AI visibility. Even when your site is not directly linked in the AI Overview, being mentioned as a source in the narrative builds entity authority. Focus on getting your brand name associated with specific topics across the web, not just accumulating links.
The term "Answer Engine Optimization" (AEO) has been floating around since 2024, but in 2026 it actually describes a distinct practice. Traditional SEO optimized pages for search engine rankings. AEO optimizes content for citation in AI-generated answers, whether those come from Google's AI Overviews, ChatGPT's web search, Perplexity, or any other generative search interface.
The tactics overlap with traditional SEO but the priorities shift. Domain authority still matters. Backlinks still help. But fact density, structural clarity, and brand entity signals now carry more weight than they used to. A page with moderate domain authority but extremely well-structured, data-rich content can get cited in an AI Overview over a higher-authority page that is poorly structured.
The sites winning at AEO in 2026 share three characteristics. First, they have a clear topical focus with consolidated content (no keyword cannibalization). Second, they use comprehensive schema markup so the AI parser can identify and extract structured data efficiently. Third, they publish proprietary content that creates primary sources rather than restating existing information.
There is a parallel here to how domain-specific language models outperform generic ones by going deep instead of wide. The same principle applies to content strategy. A site that goes extremely deep on a narrow topic, with original research and dense structural markup, will outperform a site that covers everything superficially.
Not every site is affected equally, and the recovery strategy depends on what kind of content you publish.
For B2B SaaS companies, the play is straightforward: publish benchmark data, customer case studies with real metrics, and comparison content that includes proprietary testing. Your product pages are less affected by AI Overviews (transactional intent still drives clicks). Your blog traffic on informational queries is what took the hit. Shift the blog from "explain the category" to "prove our expertise with data."
For publishers and media outlets, the situation is harder. News content is increasingly summarized in AI Overviews. The defense is speed (break stories before the AI can synthesize them) and depth (long-form investigative pieces that AI cannot compress into a paragraph without losing the point). Commodity news coverage is the most vulnerable content type on the internet right now.
For e-commerce sites, AI Overviews are eating "best X for Y" and product comparison queries. The counter-strategy involves rich product schema, unique first-hand reviews (not reworded manufacturer descriptions), and comparison tables with proprietary scoring methodology. If your content looks like something AI could generate by pulling specs from manufacturer sites, it will get replaced by exactly that.
For developer-focused sites like this one, the agentic AI architecture that powers AI Overviews is also an opportunity. Technical tutorials with runnable code, real benchmarks on specific frameworks, and opinionated architectural guides backed by production experience are the content types that still earn clicks. The AI can summarize "how to set up a Next.js project" but it cannot replicate "here is what broke when we scaled our Next.js app to 50,000 daily active users and how we fixed it."
Is the golden age of organic search traffic over? Probably, for a certain definition of "organic search traffic." If your model depended on ranking for informational queries and converting visitors from those queries, yes, that model is dying. AI Overviews are not going away. They are expanding. And competing AI search products from other companies are growing alongside them.
But search traffic is not the same as search visibility. Being cited in an AI Overview, even without a direct click, builds brand awareness. Being the authoritative source that the AI references builds the kind of trust that converts users later, through direct searches for your brand name, through word-of-mouth, through the accumulation of "I keep seeing this company mentioned whenever I research this topic."
The SEO discipline is not dead. It is splitting into two tracks. Track one optimizes for clicks where clicks still exist: transactional queries, navigational searches, long-tail topics that AI Overviews have not yet absorbed. Track two optimizes for citation and brand visibility inside AI-generated answers where clicks have disappeared. You need both.
No. Informational queries like "what is X" and "how does Y work" are most heavily affected. According to Semrush data, informational queries dropped from 89% to 57% of AI Overview triggers between 2024 and 2025, while commercial and transactional queries are increasingly getting overviews. Navigational queries (searching for a specific brand or site) are least affected.
Almost certainly not. You cannot opt out of AI Overviews specifically; you can only use noindex or nosnippet tags, which also remove you from regular search results. The visibility cost of disappearing from search entirely far outweighs the click loss from AI Overviews. The better strategy is to optimize for citation so your brand is mentioned in the overview.
Check Google Search Console for queries where your impressions remain stable but clicks have dropped. This pattern, same visibility but fewer clicks, is the signature of AI Overview impact. You can also use Semrush's Organic Rankings tool to see which of your keywords trigger AI Overviews and whether your site is cited in them.
Traditional SEO optimizes pages to rank in search results. Answer Engine Optimization (AEO) optimizes content to be cited as a source in AI-generated answers from Google, ChatGPT, Perplexity, and similar tools. AEO places higher emphasis on fact density, schema markup, brand entity signals, and content structure that is easy for AI parsers to extract and attribute.
Zero-click is structural, not temporary. AI Overviews reduce click-through rates from 15% to 8% on affected queries, per Pew Research. This is not an algorithm update you can wait out.
Keyword cannibalization now kills AI visibility. Multiple pages targeting similar keywords confuse generative engines about your authority. Consolidate overlapping content into single pillar pages with 301 redirects.
Structure content for AI extraction. Lead each section with a direct 40-to-60-word answer. Implement FAQ, HowTo, Article, and Dataset schema markup. High fact density with clear structure gets cited; fluff gets skipped.
Publish what AI cannot fake. Original benchmarks, first-party case studies, proprietary research, and named expert attribution create primary sources that AI must cite rather than synthesize around.
Build for citation, not just clicks. Brand mention signals increasingly influence AI visibility. Even without direct clicks, being cited as an authority in AI-generated answers builds compounding trust and awareness.
| ~2% |
| ~8% |
| Growing |
Redirect every retired URL to the pillar page. Update every internal link across your site that pointed to the old URLs. Broken internal links or redirect chains weaken the signal transfer.
| Step-by-step tutorials and guides |
| High. Structured steps are frequently pulled into AI summaries |
| Dataset | Pages with original data, benchmarks, statistics | Very High. Proprietary data gets cited more than synthesized opinions |
| Organization | Homepage, about page | Medium. Establishes brand entity for brand mention signals |