AI Search Is Rewriting SEO: What Site Owners Should Actually Do About It
Fewer than a third of Google searches now send a click. Here's how AI search is reshaping SEO — and the concrete moves site owners should make right now.

For twenty years, the deal between websites and search engines was simple: publish something useful, rank for the query, collect the click. That deal is now being renegotiated without your signature. Google's AI Overviews answer questions directly on the results page, AI Mode replaces the results page entirely, and a growing slice of queries never touches Google at all — they go straight to ChatGPT, Perplexity, or Claude. Understanding ai search seo — how to stay visible when the "search engine" is a language model composing an answer rather than a list of ten blue links — has become the difference between publishers who adapt and publishers who quietly watch their traffic charts sag.
The data is not subtle. According to SparkToro's 2026 analysis, fewer than one in three Google searches now sends a click to the open web. That's not a dip; it's a structural change in how the internet distributes attention.
The good news: this is not the death of SEO. It's a fork in it. The sites that win in AI search look a lot like the sites that won in classic search — with a few new habits, and a few hard decisions about who gets to crawl your content and on what terms.
The numbers behind the panic
Start with the scale of the shift. Zero-click search data for 2026 shows AI Overviews now appearing on about 8% of all Google queries — over a third of informational ones — and when one is present, organic click-through rates fall sharply: the same data puts top-result click-through at 19.8% with an AI Overview versus 31.7% without. Pew Research found the same pattern, with users clicking a traditional result on roughly 8% of visits with an AI summary versus 15% without. Google's AI Mode is more drastic still: because it replaces organic results with a conversational interface, around 88% of AI Mode sessions end without a click to any website. Gartner, meanwhile, projected that traditional search volume would decline 25% by 2026 as queries migrate to conversational interfaces — a forecast that no longer looks aggressive.
There is a genuine silver lining buried in this. The clicks that survive are better clicks. Users who read an AI summary and still click through have already qualified themselves — several 2026 analyses report that these visitors convert meaningfully better than the pre-AI average, and referrals from ChatGPT and Perplexity convert at multiples of classic organic traffic. You're getting fewer visitors, but a higher share of them actually want what you're selling.
AI search SEO is mostly still SEO — with one big twist
Google's own guidance says optimizing for its generative features "is still SEO," and it's largely right. AI Overviews and retrieval-based engines like Perplexity and ChatGPT's search mode don't conjure answers from nothing; they fetch live web pages, rank them, and synthesize. If your page can't rank, it can't get cited. Crawlability, site speed, clear information architecture, and genuine expertise still matter exactly as much as they did in 2023.
The twist is the goal. Classic SEO optimized for a click; generative engine optimization (GEO) optimizes to be quoted. The industry has coalesced around this as an additive layer on top of SEO, not a replacement for it. If you want to see how differently these engines behave, run your own money queries through Perplexity — its citation behavior is unusually transparent, which makes it a useful diagnostic tool.
What actually gets cited by AI engines
Across the platforms, the pages that earn citations share recognizable traits. The practical checklist:
- Front-load the answer. Retrieval-based engines judge relevance heavily on a page's opening content. Your first 200 words should answer the query directly — not wind up to it.
- Structure for extraction. Clear H2/H3 headings, question-shaped subheads, lists, and tables give a model clean blocks to lift. Walls of prose get skipped.
- Add schema markup. FAQPage and Article structured data in JSON-LD make your Q&A pairs machine-readable — each one becomes an explicit citation candidate.
- Publish something only you can publish. Original data, first-hand testing, named expertise. Models are trained to prefer sources that look authoritative, and commodity content is precisely what they synthesize away.
- Earn third-party mentions. AI engines weigh how often your brand appears across credible sources. Digital PR just became an AI visibility tactic.
- Keep it fresh. Perplexity in particular favors recently updated pages for time-sensitive queries. Stale content decays faster in AI search than it did in classic rankings.
The single most important shift: stop optimizing to win a click and start optimizing to be the source the AI quotes. Your brand can now be seen by thousands of people in answers your analytics will never record — and being absent from those answers is the new page two.
Decide who crawls you — and whether they pay
Visibility is only half the equation. The other half is leverage, and 2026 finally gave site owners some. Cloudflare's pay-per-crawl program lets domain owners set a per-request price for AI crawlers instead of the old binary allow/block choice. And as TechCrunch reported, from September 15, 2026, Cloudflare will block "mixed-use" crawlers by default on ad-supported pages, while separating bots into categories: training and agent crawlers blocked by default for new domains, search crawlers still allowed.
That category split is the strategy in miniature. Blocking training crawlers protects your content from being absorbed into future models without compensation. Blocking search crawlers removes you from AI answers entirely — which, given where users are going, is usually self-harm. Meanwhile, the emerging llms.txt standard (adopted by roughly 10% of domains so far) lets you point AI systems at your cleanest, highest-value pages, the way sitemaps once did for Googlebot. It's early and unevenly honored, but it costs an afternoon to implement.
Measure what's left — and what's new
Your old dashboard now lies to you. Impressions without clicks look like failure but may represent an AI Overview citation reaching thousands of readers. Rebuild your measurement around three things: referral traffic from AI domains (chatgpt.com, perplexity.ai, and friends — segment them in analytics), conversion rate rather than raw sessions, and brand presence inside AI answers. That last one you have to check manually or with tooling: ask the major assistants the questions your customers ask, and see whether you're in the answer. A paid ChatGPT account is worth it for this alone, since search behavior differs between tiers.
Then set expectations honestly. Informational traffic — the "what is X" and "how does Y work" queries — is largely gone and not coming back. Transactional and high-intent traffic is holding up far better; notably, AI Overview coverage of e-commerce queries actually dropped in Google's 2026 data. Invest where a click still happens: comparisons, tools, original research, and content tied to a purchase decision.
FAQ
Is SEO dead now that AI answers most queries?
No, but it's smaller and different. AI engines source their answers from pages that already rank well, so classic SEO fundamentals remain the entry ticket. What's dead is the assumption that visibility automatically equals traffic — a growing share of your "wins" now happen inside answers, as citations and brand mentions, rather than as sessions in your analytics.
Should I block AI crawlers from my site?
Split the decision by crawler type. Blocking training crawlers protects your content from uncompensated model training and rarely costs you visibility today. Blocking search and retrieval crawlers removes you from AI answers, which is where your audience increasingly is. Tools like Cloudflare's AI Crawl Control and pay-per-crawl let you enforce that distinction — and potentially get paid for access instead of choosing between all or nothing.
What is GEO, and do I need a separate strategy for it?
Generative engine optimization is the practice of getting your content cited inside AI-generated answers rather than ranked in a list of links. It's best treated as a layer on top of SEO, not a replacement: front-loaded answers, extractable structure, schema markup, and third-party authority signals. If your SEO fundamentals are weak, no amount of GEO tactics will get you cited.
About the author
UbedullaFounder & Editor
Founder and editor of The Bot Post, covering AI news and technology.


