SEO 2.0: how AI search and assistants changed SEO, layer by layer
SEO 2.0 is search optimization for a web where AI assistants fetch pages and quote short passages. The base is unchanged: crawlable, indexed, trustworthy pages. On top come answers in the first lines (AEO), being named and cited by assistants (GEO), pages that are easy to extract (AIO) and pages that satisfy the visitor (SXO).
SEO 2.0 is not a new rulebook. Search engines still crawl, index and rank pages. What changed is who reads your page first and how little of it they pass on: an AI crawler that may not run your scripts, an assistant that quotes two sentences and links a few sources, a firewall that decides which bots get in. Below are five layers, the checks for each and a short checklist.
On this page
- What did old SEO look like, and why does it fail now?
- What changed when AI search and assistants arrived?
- What are the five layers of SEO 2.0?
- Layer 1, the SEO base: can crawlers reach, read and trust your page?
- Layer 2, AEO: does your page answer the question in two lines?
- Layer 3, GEO: do AI assistants name and cite you?
- Layer 4, AIO: is your page easy for a machine to take apart?
- Layer 5, SXO: does the visitor who clicks get what they came for?
- What should you check first?
- How can you check this with AISeeWhy?
What did old SEO look like, and why does it fail now?
The old playbook had four habits: repeat the keyword, get links from anywhere, publish as many pages as possible and chase a position on page one. Google's spam policies now name the first three as violations: keyword stuffing, buying or selling links for ranking purposes, and scaled content abuse, meaning many pages made mainly to manipulate rankings.
Position still matters, but it measures less. Google says an AI Overview occupies a single position and all of its links share it. In ChatGPT, Claude, Perplexity or Gemini there is no page one at all: an answer cites a handful of sources, and your page is either among them or not.
- Stuff keywords → answer the question in plain words.
- Buy any links → be described accurately on sites people trust.
- Pump out pages → one useful page per topic.
- Chase page one → be the source an answer quotes, and measure that too.
What changed when AI search and assistants arrived?
Three mechanics explain most of SEO 2.0. First, AI systems read your site through their own crawlers: OAI-SearchBot and ChatGPT-User for ChatGPT, Claude-SearchBot and Claude-User for Claude, PerplexityBot and Perplexity-User for Perplexity, and Googlebot for AI Overviews and AI Mode. Google renders JavaScript, but OpenAI, Anthropic and Perplexity do not say their agents execute it, so text added by scripts may never reach them.
Second, an assistant answers in its own words and links a few sources. A passage that names its subject and makes sense alone is easier to use. Google says you do not need to chop pages into small pieces for its AI features; clear writing for people is enough.
Third, access can fail silently. A robots.txt group that names GPTBot or ClaudeBot can catch more than intended, and a CDN or bot protection can answer a crawler with a 403 or a challenge page while your browser sees the site normally. Anthropic says its bots do not try to bypass CAPTCHAs. In our study of the 1,000 most popular domains, 19.5% of the 518 sites that served a robots.txt blocked at least one AI search or user crawler from the home page.
What are the five layers of SEO 2.0?
Each layer depends on the ones below it: a perfect answer block does nothing if the crawler gets a challenge page.
- SEO base: pages crawlers can reach and read, linked clearly, indexed and checked in Search Console.
- AEO, answer engine optimization: the answer in the first two lines, FAQs on key pages, a complete Business Profile for local businesses.
- GEO, generative engine optimization: being named, cited and quoted by ChatGPT, Gemini, Perplexity and Claude, including through other sites, and tracking it.
- AIO, AI optimization: one topic per page, lists and tables, prices and specs written as text.
- SXO, search experience optimization: the page matches the search, works well on a phone, shows proof and has one clear next step.
Layer 1, the SEO base: can crawlers reach, read and trust your page?
Every other layer assumes this one. Run these checks on your home page and the three pages that earn you the most.
Read every group in /robots.txt, not only User-agent: *, and allow the AI search and user crawlers you want to be cited by. Training crawlers such as GPTBot and ClaudeBot are separate tokens you can block without leaving AI search. Then fetch the page with a crawler's user agent and search the raw HTML for a sentence from the page; a 403, a 429 or a count of 0 needs a closer look.
curl -s -o /dev/null -w "%{http_code}\n" -A "Mozilla/5.0 (compatible; OAI-SearchBot/1.4; +https://openai.com/searchbot)" https://example.com/page/
curl -s https://example.com/page/ | grep -c "a sentence from the page" Some firewalls also check the IP address, which a test from your machine cannot reproduce. OpenAI, Anthropic and Perplexity publish their crawler IP ranges; allow those rather than trusting the user agent alone.
Link key pages with ordinary <a href> links and descriptive anchor text: Google says it can generally only crawl links that are <a> elements with an href, and discourages text like "click here". Keep a sitemap of the pages you want found.
In Google Search Console, URL Inspection shows whether a page is indexed. Look for a stray noindex, a nosnippet or a tiny max-snippet: Google says nosnippet also stops content from being used as a direct input for AI Overviews and AI Mode. The Generative AI performance report shows how often links to your pages appeared in Google's AI features.
Layer 2, AEO: does your page answer the question in two lines?
Put a direct answer of two or three sentences, roughly 40 to 60 words, right under the H1, and explain after it. It should make sense when quoted alone: name the subject instead of writing "it", and include the one number or condition that matters.
Use the questions people ask as headings: "How much does it cost per user?" says what the section answers, "Our pricing philosophy" does not. On pricing, product and service pages, add a short visible FAQ with the questions buyers ask before deciding. FAQ rich results stopped appearing in Google on May 7, 2026, so FAQPage markup no longer changes how a result looks; it is valid and harmless, but the visible text is what readers and crawlers use.
If customers visit you in person, keep your Google Business Profile complete and consistent with your site. Google says Business Profile information can make local businesses visible in its AI responses.
Layer 3, GEO: do AI assistants name and cite you?
GEO asks three questions: is your brand named, is your page cited, is your wording quoted? An assistant can name you without linking you, or cite you while recommending someone else, so track them separately.
Much of what assistants say about a business comes from other sites: reviews, directories, comparison articles and forums such as Reddit. OpenAI says ChatGPT considers third-party content when it chooses products. Make sure existing listings describe you correctly. In forums, answer the question and say who you are. Google says seeking inauthentic mentions is not as helpful as it may seem, and fake reviews break most platforms' rules.
To track it, ask 10 to 30 buyer questions in each assistant on a fixed schedule and record whether you are named, where in the list and which sources are cited. Answers vary between runs, so watch the trend. For traffic, Google Analytics 4 has had an AI Assistant channel since 13 May 2026, and Bing Webmaster Tools has an AI Performance report for Copilot.
Layer 4, AIO: is your page easy for a machine to take apart?
A page about one topic is easier to match to a question than one that covers five. If a page answers two unrelated questions, split it and link the two pages.
Put comparisons in an HTML <table> and steps in a numbered list, not in images or sliders. Write prices, specs, limits and opening hours as text with units and currency: "from $19 per user a month, billed yearly" can be quoted, a price inside a calculator cannot. Google lists keeping important content in text form among its best practices for AI features.
Add JSON-LD that matches the visible text: Organization with sameAs links to your official profiles, and Article, Product or SoftwareApplication per page, with dateModified equal to the visible updated date. Google says no special schema is needed for its AI features, so treat markup as a clear label, not a ranking switch.
Layer 5, SXO: does the visitor who clicks get what they came for?
Start with intent: search your main query and see what ranks. If the results are guides and your page is a sales page, it does not match the search, whatever its keywords say.
Then test on a phone. Google uses the mobile version of a page for indexing and ranking. Check the viewport meta tag and Core Web Vitals in Search Console; Google's good thresholds are LCP within 2.5 seconds, INP under 200 milliseconds and CLS under 0.1.
Visitors from an AI answer have often read a summary already and come to confirm and act. Show real reviews with dates and sources, results you can back up and who stands behind the product; Google's helpful content guidance calls trust the most important part of E-E-A-T. End each key page with one clear call to action.
What should you check first?
This list follows the stack from the bottom. Most items take a few minutes each on one page.
- robots.txt allows the AI search and user crawlers you want, in every group that names them.
- Your CDN or firewall gives AI crawlers the same page and status code as a browser.
- The main text is in the server's HTML, without JavaScript.
- Key pages are indexed, with no stray
noindex,nosnippetor tinymax-snippet, and linked with descriptive anchors. - Each key page answers its question in the first two or three sentences, under question headings.
- Pricing and product pages have a short visible FAQ and state prices and specs as text and tables.
- JSON-LD matches the page, and the visible date matches
dateModified. - The page works on a phone, shows proof and ends with one call to action.
- Other sites describe you correctly, and you track AI answers for 10 to 30 buyer questions.
How can you check this with AISeeWhy?
The free check covers 16 AI and search crawlers: it fetches your page with each crawler's user agent next to a real browser from the same address, applies your robots.txt to each crawler (tokens that exist only in robots.txt, such as Google-Extended, are checked there), spots firewall blocks and challenge pages, and measures how much text arrives without JavaScript. The free preview shows how each crawler fared and the most serious finding in full.
The full report, on a paid plan or as a one-time report, lists every finding with a fix: the short answer under the title, question headings, links to sources, JSON-LD (Article fields, publisher, sameAs, FAQ), noindex, nosnippet and max-snippet, canonical, title, description, viewport and slow responses, plus the sitemap, llms.txt, hreflang and whether dateModified agrees with the sitemap.
Paid plans add monitoring that emails you when a page's result gets worse, a site audit for broken links, orphan and thin pages and duplicate titles, and AI visibility: it asks ChatGPT, Claude, Perplexity and Gemini your prompts through their APIs with web search and records whether you are named, your place in the list and which sites are cited. API answers can differ from the apps, and it does not cover Google AI Overviews. You can also connect Search Console read-only to turn real queries into prompts, and upload a server log to see which AI crawlers really fetch your pages.
No tool can tell you whether your content is original and useful, or what other sites say about you. Those parts stay with you.
Fixes for this problem
- OAI-SearchBot is blocked
- Page content needs JavaScript
- No clear next step on a commercial page
- No short answer under the H1
Next steps
- robots.txt tester and checker for AI crawlers
- llms.txt checker and validator
- See an example full report
- AI crawler monitoring
- Plans and prices
Crawlers in this article
Frequently asked questions
Is SEO 2.0 a replacement for SEO?
No. Google says a page must be indexed and eligible for a snippet to appear in AI Overviews and AI Mode, and AI search crawlers need a crawlable page with text in the HTML. SEO 2.0 adds layers on top of classic SEO; it does not remove the base.
What is the difference between AEO and GEO?
AEO is about your own page: a direct answer in the first lines, question headings and FAQs, so any answer engine can use it. GEO is about the result in generative assistants: whether ChatGPT, Gemini, Perplexity or Claude name your brand and cite your pages, which also depends on what other sites say about you.
Do I need llms.txt for SEO 2.0?
No. Google says its Search does not use llms.txt, and OpenAI, Anthropic and Perplexity do not list it as a requirement for their crawlers. It is optional and harmless. Crawler access, text in the HTML and clear answers matter far more.
Does FAQ schema still help?
Not visibly in Google: FAQ rich results stopped appearing on May 7, 2026. FAQPage markup is still valid and harmless, but no search or AI engine documents that it earns citations. Keep the visible FAQ, because that text is what readers and crawlers use.
Sources
- Google Search Central: Spam policies for Google web search
- Google Search Console Help: What are impressions, position, and clicks?
- Google Search Central: AI features and your website
- Google Search Central: Optimizing your website for generative AI features on Google Search
- OpenAI: Overview of OpenAI crawlers
- Anthropic: Does Anthropic crawl data from the web, and how can site owners block the crawler?
- Perplexity: Perplexity crawlers
- Google Search Central: JavaScript SEO basics
- Google Search Central: Link best practices for Google
- Google Search Console Help: URL Inspection tool
- Google Search Central: Robots meta tag, data-nosnippet and X-Robots-Tag
- Google Search Console Help: Generative AI performance report (Search)
- Google Search Central Blog: Changes to HowTo and FAQ rich results (August 2023)
- Google Search Central: Latest Google Search documentation updates (FAQ deprecation and removal)
- Schema.org: FAQPage
- OpenAI: Shopping with ChatGPT Search
- Google Analytics Help: What's new in Google Analytics (13 May 2026, New AI Assistant traffic measurement)
- Microsoft Bing: Introducing AI Performance in Bing Webmaster Tools Public Preview
- Google Search Central: Organization structured data
- Google Search Central: Mobile-first indexing best practices
- Google Search Central: Understanding Core Web Vitals and Google search results
- Google Search Central: Creating helpful, reliable, people-first content
- AISeeWhy: AI Crawler Access Report, 2026-11