AI SEO Optimization

AI SEO Optimization Skill

Comprehensive guide for optimizing Next.js sites for Google's generative AI features (AI Overviews, AI Mode) and emerging browser agent interactions.

Framework note: Although examples use Next.js conventions (e.g., generateMetadata), all checklist items apply to any web framework. For non-Next.js sites, substitute framework-equivalent SSR and metadata APIs.

Source: Google AI Optimization Guide Companion: Build Agent-Friendly Websites Verify current guidance: These references reflect guidance as of the skill's authoring date. Always verify current Google documentation before advising clients.

Core Principle

AI Search optimization IS SEO. There is no separate "AEO" or "GEO" discipline. Google's AI features use the same Search index, the same ranking signals, and the same content quality assessments. If your page is well-optimized for Search, it's well-optimized for AI features.

1. How Google's AI Features Discover Content

Google's generative AI features use Retrieval-Augmented Generation (RAG):

1. User query triggers query fan-out (multiple sub-queries)
2. Each sub-query retrieves pages from the existing Search index
3. AI synthesizes answers from retrieved content
4. Citations link back to source pages

Requirements for inclusion:

• Page MUST be indexed (verify in Search Console)
• Page MUST allow snippets (no nosnippet meta directive)
• Content MUST be server-rendered (not client-only JavaScript)
• Content MUST be accessible to Googlebot (no login walls for indexed content)

2. Agent-Friendly UX Audit Checklist

Browser agents (Google's Mariner, OpenAI's Operator, etc.) interact with sites through three channels:

1. Screenshots — Vision model identifies elements visually
2. HTML/DOM — Understands nesting, hierarchy, relationships
3. Accessibility tree — Roles, names, states of interactive elements

Semantic HTML

• All clickable elements use <button> or <a href> (NEVER <div onClick> or <span onClick>)
• All form inputs have associated <label htmlFor="id"> (not just placeholder text)
• Heading hierarchy is logical (h1 > h2 > h3, no skips)
• Lists use <ul>/<ol>/<li> (not styled divs)
• Tables use <table> with <th> headers for tabular data
• Navigation uses <nav> with proper aria-label
• Main content wrapped in <main>
• Sections use <section> or <article> with headings

Accessibility Tree

• All interactive elements have accessible names (visible text, aria-label, or aria-labelledby)
• ARIA roles used ONLY when semantic HTML isn't available
• Skip-to-content link is first focusable element
• Focus order matches visual order (tabindex not misused)
• Modal dialogs use <dialog> or proper role="dialog" with focus trapping
• Dynamic content changes announced via aria-live regions

Visual Stability

• No layout shifts after load (CLS < 0.1)
• Interactive elements have stable positions (no jumping during scroll)
• No transparent overlays blocking clickable elements (ghost overlays)
• All actionable elements meet minimum target size of 24×24 CSS px (WCAG 2.2 SC 2.5.8)
• Images have explicit width and height (prevent layout shift)

CSS Signals for Agents

• cursor: pointer on all clickable elements (strong actionability signal)
• Focus indicators visible on keyboard navigation (:focus-visible styles)
• Interactive states are distinct (:hover, :active, :focus)
• Disabled elements have pointer-events: none and opacity reduction
• No display: none on elements that should be accessible to screen readers

Form Accessibility (Critical for Agent Interactions)

• Every <input> has an associated <label> with htmlFor
• Required fields marked with aria-required="true" or required attribute
• Error messages linked with aria-describedby
• Form groups use <fieldset> with <legend>
• Submit buttons have clear text (not just icons)
• Autocomplete attributes set for common fields (name, email, tel, address-*)

3. E-E-A-T Assessment (Critical for YMYL Sites)

For sites dealing with health, finance, safety, or life decisions (Your Money or Your Life), E-E-A-T signals are critical for AI citation.

Experience

• Content demonstrates first-hand knowledge (not aggregated summaries)
• Real user reviews and testimonials with attribution
• Case studies or examples from actual experience
• Visual evidence (photos, videos) of real experiences

Expertise

• Author credentials displayed clearly (degrees, certifications, years of experience)
• Author pages exist with Person schema and knowsAbout properties
• Content reviewed by subject matter experts (editorial review notice)
• Specialized terminology used correctly with explanations for lay readers

Authoritativeness

• About page with team credentials, methodology, and mission
• Clear data source disclosures ("Where our data comes from")
• "How we rate" or methodology explanation pages
• Institutional authority signals (partnerships, certifications, media mentions)
• Consistent NAP (Name, Address, Phone) across web presence

Trustworthiness

• Contact information easily findable (not buried in footer only)
• Privacy policy and terms of service accessible from all pages
• Transparent pricing when applicable (no hidden fees)
• Content is accurate and up-to-date (publish/update dates visible)
• Corrections policy or update log for changed information
• HTTPS enforced, security headers properly configured

4. Non-Commodity Content Criteria

Google's AI features prefer content that provides unique value. Assess content against these criteria:

Non-Commodity Content (Google Favors)

• Unique data analysis from proprietary sources (not publicly available datasets)
• First-hand experiences and detailed reviews with specific details
• Expert-led insights that go beyond commonly available information
• Original research with methodology and reproducible findings
• Interactive tools and calculators with unique logic
• Comparison data that requires effort to compile (pricing, availability)
• Local knowledge that can't be found without physical presence

Commodity Content (Avoid or Upgrade)

• ❌ "X Tips for [topic]" without unique insights (generic listicles)
• ❌ Summaries of information freely available elsewhere
• ❌ Generic advice easily produced by any AI model
• ❌ Content created primarily for keyword targeting
• ❌ Thin pages with minimal original value
• ❌ Automatically generated content without expert review

Content Upgrade Strategy

For existing commodity content, improve by adding:

1. Unique data points from proprietary sources
2. Expert commentary with attribution
3. Real examples with specifics (names, dates, outcomes)
4. Interactive elements (calculators, comparison tools)
5. Original media (photographs, diagrams, video)

5. Technical SEO for AI Discovery

Indexing & Crawlability

• All key pages indexed (verify in Google Search Console)
• No accidental noindex directives on important pages
• No nosnippet or data-nosnippet blocking content extraction
• robots.txt doesn't block critical resources
• Crawl budget optimized (no low-value pages wasting crawl resources)
• XML sitemap includes all valuable pages (updated automatically)

Server-Side Rendering

• Critical content renders on server (not client-only JavaScript)
• Dynamic content accessible without user interaction
• No infinite scroll hiding content from crawlers
• Metadata generated server-side (generateMetadata in Next.js)

URL & Content Structure

• Canonical tags correctly set (prevent duplicate content confusion)
• Internal linking connects related content (topic clusters)
• Breadcrumb navigation with schema markup
• Clean URL structure reflecting content hierarchy
• Pagination handled with view-all option or clear next/previous links (note: rel="next"/rel="prev" is no longer used as a Google indexing signal)

Performance & Page Experience

• Core Web Vitals in "Good" range (LCP < 2.5s, INP < 200ms, CLS < 0.1)
• HTTPS enforced with valid certificate
• No intrusive interstitials blocking content
• Mobile-friendly (responsive design)
• Font loading doesn't cause layout shift (display: swap)

6. Structured Data Patterns

Structured data helps Google understand content relationships. While NOT required for AI features, it improves overall search presence.

LocalBusiness (Community/Location Pages)

{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Community Name",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "123 Main St",
    "addressLocality": "City",
    "addressRegion": "FL",
    "postalCode": "33101"
  },
  "geo": { "@type": "GeoCoordinates", "latitude": 25.76, "longitude": -80.19 },
  "telephone": "+1-555-0100",
  "priceRange": "$$-$$$",
  "openingHours": "Mo-Su 00:00-24:00",
  "aggregateRating": { "@type": "AggregateRating", "ratingValue": "4.5", "reviewCount": "28" },
  "sameAs": ["https://g.page/community", "https://facebook.com/community"]
}

Person (Author/Expert Pages)

{
  "@context": "https://schema.org",
  "@type": "Person",
  "name": "Author Name",
  "jobTitle": "Senior Care Advisor",
  "knowsAbout": ["Assisted Living", "Memory Care", "VA Benefits"],
  "sameAs": ["https://linkedin.com/in/author"],
  "worksFor": { "@type": "Organization", "name": "Company Name" }
}

Article (Blog/Resource Posts)

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Article Title",
  "author": { "@type": "Person", "name": "Author", "url": "/about/authors/slug" },
  "datePublished": "2026-01-15",
  "dateModified": "2026-05-01",
  "publisher": { "@type": "Organization", "name": "Site Name" }
}

FAQPage (FAQ Sections)

{
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is the cost of assisted living?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "The average cost varies by state..."
      }
    }
  ]
}

7. What NOT to Do (Mythbusting)

These are explicitly confirmed as unnecessary or harmful by Google:

8. Implementation Priorities

Quick Wins (< 1 day each)

1. Add cursor: pointer CSS for all interactive elements
2. Add skip-to-content link to root layout
3. Verify no nosnippet directives on key pages
4. Check all forms have proper <label> associations
5. Verify heading hierarchy on top landing pages

Medium Effort (1-3 days each)

1. Semantic HTML audit of all components (replace <div onClick>)
2. Author/expert page creation with Person schema
3. "About Our Data" / methodology page
4. Image alt text quality audit and improvement
5. Internal linking optimization

High Effort (1+ week each)

1. Non-commodity content strategy and execution
2. Interactive tools (calculators, comparison widgets)
3. Comprehensive accessibility audit and remediation
4. E-E-A-T content enrichment across all page types
5. Video/media content creation

9. Audit Workflow

Before beginning the audit, confirm you have access to: (1) the site URL or representative page HTML, (2) Google Search Console data if available, and (3) the page types to audit (home, landing, blog, etc.).

If no site URL, HTML, or page content is provided, respond: "To perform this audit, please share the site URL, representative page HTML, or specific components you want evaluated."

Representative HTML alone cannot support checks that need live signals — index status, live crawl/HTTP accessibility, Core Web Vitals field data, and Search Console signals. When the corresponding source is not provided (no live URL, no CrUX/PageSpeed access, no Search Console), report those items as "Not verifiable — requires <live URL | CrUX/PSI | Search Console>" rather than inferring a pass/fail from static markup.

When asked to audit a site for AI optimization, follow this order:

1. Technical baseline — Verify indexing, SSR, canonical tags, Core Web Vitals
2. Agent-friendly audit — Run through Section 2 checklist on key templates
3. E-E-A-T assessment — Evaluate Section 3 for YMYL compliance
4. Content quality — Assess top pages against Section 4 criteria
5. Structured data — Validate existing schema, identify gaps
6. Recommendations — Prioritize by impact/effort ratio

10. Monitoring & Measurement

References

• Google AI Optimization Guide
• Build Agent-Friendly Websites
• Creating Helpful, Reliable, People-First Content
• Google Search Quality Rater Guidelines
• Universal Commerce Protocol (UCP)