This is the most commonly failed check — and the most consequential. Many sites inadvertently block AI crawlers through overly broad robots.txt rules that were written before AI bots existed. A single line like User-agent: * Disallow: / blocks every bot including all AI crawlers. So does blocking User-agent: GPTBot explicitly — which became common in 2023 when publishers panicked about AI scraping.

1. Open your browser and go to yourdomain.com/robots.txt. Read every rule carefully.
2. Search for these bot names: GPTBot, ClaudeBot, PerplexityBot, anthropic-ai, CCBot. If any are followed by Disallow: / — that crawler is blocked.
3. Check for wildcard blocks: User-agent: * with Disallow: / blocks every bot. Fix by either removing the rule or explicitly re-allowing AI crawlers beneath it.
4. Verify your llms.txt file exists at yourdomain.com/llms.txt. This file guides AI crawlers to your most authoritative content.

Indexing is the prerequisite for everything else. A page that isn’t indexed by Google cannot appear in organic results, cannot be cited in Google AI Overviews, and in many cases cannot be retrieved by AI search systems that rely on Google’s index. An Ahrefs study confirmed that 76.1% of Google AI Overview citations (mid-2025) came from pages already in the organic top 10 — which means indexed, ranking pages.

“Standard Search Essentials apply to AI Overview inclusion. There is no special markup to qualify — inclusion requires indexation and snippet eligibility.”

— 201 Creative GEO Audit Checklist, April 2026

1. In Google Search Console, go to Pages → Why pages aren’t indexed. Address any “Discovered but not indexed,” “Crawled but not indexed,” or “noindex” errors on important pages.
2. Run a site search in Google: site:yourdomain.com. The number of results gives a rough indexed page count. Compare to your actual sitemap count.
3. Check your XML sitemap at yourdomain.com/sitemap.xml. Confirm it includes all important pages and is submitted in Search Console.
4. Identify any critical pages accidentally tagged with <meta name="robots" content="noindex"> — common on WordPress sites after a staging environment is pushed live.

Pages loading in under 1.5 seconds receive 3× more traffic than slower pages. Beyond ranking, page experience directly affects AI Overview inclusion — Google’s AI systems favor snippet-eligible pages, and poor Core Web Vitals can reduce snippet eligibility even for well-optimized content.

1. Run your 5 most important pages through PageSpeed Insights. Check both Mobile and Desktop scores.
2. In Google Search Console → Experience → Core Web Vitals, check how many URLs are in “Poor” or “Needs Improvement” status.
3. Fix quick wins first: compress images (use WebP format), enable browser caching, minimize CSS/JS, and use a CDN if your server is far from your audience.

Structured data (JSON-LD schema markup) is the single most high-leverage LLMO action you can take. LLMs use structured data as a “cheat sheet” to understand your content without parsing raw HTML. Optimal.dev’s 2026 LLMO audits found that sites with comprehensive structured data are retrieved 2.8× more frequently by RAG-based AI systems than sites without it.

Most sites have partial schema at best — Organization or BreadcrumbList on the homepage, nothing on content pages. The goal is comprehensive coverage across all page types.

1. Run your homepage, a blog post, and a service page through Google’s Rich Results Test. Note every error and warning.
2. Check for FAQPage schema on your top 10 content pages. This is the most commonly missing schema type and has the highest AIO citation impact.
3. Verify your Organization schema includes: name, url, logo, sameAs (links to LinkedIn, Twitter, Wikipedia if applicable), and contactPoint.
4. Add Person schema to every named author with name, url, jobTitle, and sameAs linking to their professional profiles.

This is the GEO check that most sites fail — not because it’s hard, but because most content was written for human readers who expect context-building introductions, not for AI extraction systems that scan for the most direct answer to the query. Google’s AIO system and AI chatbots like Perplexity both prioritize the first substantive answer they find on a page.

The fix is structural, not a complete rewrite. For every important page, locate the primary query it targets. Then ensure the first full paragraph — within 80 words — provides a clear, direct, standalone answer to that query. Longer explanations, context, and depth can follow. This mirrors the inverted pyramid used in journalism: answer first, elaborate second.

1. List your 10 most important pages and write down the primary query each targets (e.g. “What is generative engine optimization?”).
2. Open each page and read the first paragraph. Ask: does it directly answer the query within 80 words, as a standalone statement? If not, rewrite it.
3. Add a Q&A section (and FAQPage schema) to each page with 3–5 questions users would actually ask, answered in 40–80 words each.
4. Check heading formatting: headings should be phrased as questions where possible (e.g. “What Is Universal Search Optimization?” not “About Universal Search Optimization”).

E-E-A-T is not a ranking factor you can fake — it’s the sum of verifiable signals that demonstrate your content comes from real experts with real experience. The “Experience” component, added in December 2022, specifically rewards first-hand knowledge: personal accounts, case studies, and content that reflects direct involvement with the subject matter.

1. Author pages: Every piece of content should have a named author with a dedicated bio page listing credentials, experience, and external profiles (LinkedIn, publications). Anonymous or byline-free content scores poorly on E-E-A-T.
2. About page audit: Your About page should clearly state who runs the site, what their expertise is, when the organization was founded, and how to contact them. Missing or thin About pages are a significant E-E-A-T red flag.
3. External citations: Count how many external authoritative sources link to or mention your site. Tools like Ahrefs or Semrush show your referring domain count. Aim for citations from industry publications, not just general directories.
4. First-hand experience signals: Add original data, personal case studies, or first-person accounts of the topic. AI systems weight exclusive, verifiable sources more heavily than aggregated information (GEO CORE-EEAT: E01).

Ranking a single page is harder than owning a topic. Search engines in 2026 evaluate topical authority across clusters of interconnected content — not individual pages in isolation. A site that publishes multiple related articles around one theme, properly interlinked, sends stronger relevance signals than isolated high-quality pages.

This structure also directly benefits GEO performance: AI systems that use RAG retrieval pull surrounding context along with relevant passages. A well-interlinked topic cluster means AI systems retrieve richer, more authoritative context alongside your key content — making citations more likely and more accurate.

1. Map your content: list all pages grouped by topic area. Identify your “pillar” pages (comprehensive overviews) and “cluster” pages (specific subtopics).
2. Audit internal links: every cluster page should link to its pillar, and every pillar should link to all cluster pages. Use Screaming Frog to crawl your site and spot orphaned pages (pages with no internal links pointing to them).
3. Check anchor text: internal links should use descriptive, keyword-relevant anchor text — not “click here” or “read more.”
4. Identify content gaps: topics your competitors cover that you don’t. These gaps represent both SEO opportunities and GEO citation opportunities.

Google AI Overviews now appear on approximately 48% of Google queries — up from 31% in February 2025 (ALM Corp, March 2026). But their impact is not uniform: informational head terms face CTR drops of 34%–64%, while commercial and branded queries show smaller impact or even gains. You cannot manage AIO impact without knowing which of your pages are affected.

1. In Google Search Console, go to Search Results. Click Search Type → select Web. Look for queries where impressions are high but CTR is unusually low (under 1%) — these are likely AIO-affected queries.
2. Manually search your top 20 target queries in Google. Note which ones show an AI Overview. Does your site appear in the AIO? If not, why not?
3. For pages losing clicks to AIO: add direct-answer formatting (Check 5), implement FAQPage schema (Check 4), and target conversational long-tail variants of the same query.
4. Set up a simple tracking sheet: list your top 20 queries, whether AIO appears, and whether you’re cited in AIO. Update monthly.

The llms.txt convention was proposed in 2024 and saw rapid adoption in 2025. It is not a Google ranking factor, but it is used by ChatGPT Browse, Perplexity, Claude, and other AI retrieval systems to prioritize which pages to surface when answering questions related to your domain. Sites with a well-formatted llms.txt file give AI systems a direct signal about their expertise areas and most citable content.

1. Check yourdomain.com/llms.txt in your browser. If it returns a 404, you don’t have one.
2. Create llms.txt at your site root. Structure it with: a brief description of your site, a list of key topic areas, and links to your most authoritative pages.
3. Optionally create llms-full.txt containing the full text of your 5–10 most important pages, formatted in clean Markdown.
4. Verify your sitemap is linked from llms.txt so AI crawlers can discover your full content inventory efficiently.

LLMO goes beyond technical schema markup into entity clarity — the degree to which AI systems accurately comprehend and represent your brand. A brand can appear in AI-generated responses and still be misrepresented: associated with wrong products, outdated descriptions, or incorrect expertise areas. Entity disambiguation prevents this.

AI models organize their understanding of the world around entities — specific people, organizations, products, and concepts — and their relationships. Your job is to make your entity as unambiguous and well-defined as possible across all the signals AI systems use.

1. Name consistency: Search your brand name across your website, social profiles, Google Business Profile, LinkedIn, and Wikipedia (if applicable). The name should be identical everywhere — not “SearchUniversal” on some pages and “Search Universal” on others.
2. Organization schema sameAs: Add a sameAs array to your Organization JSON-LD linking to every external profile (LinkedIn, Twitter/X, Crunchbase, Wikipedia, industry directories). This creates a machine-readable identity graph.
3. About page test: Ask ChatGPT or Perplexity: “What is [your brand name]?” If the description is wrong, outdated, or missing entirely, your entity clarity is insufficient. Fix it by publishing clear, factual About content and building external citations.
4. Cross-platform presence: Being mentioned on Reddit, LinkedIn, GitHub, industry forums, and YouTube directly feeds AI training data and retrieval indexes. Build brand mentions on platforms where your audience already discusses your topic area.