LLM Best Practices

SEO & AI Search Optimization

6 min read

Search and agent discoverability as one coherent surface. Start at the foundation (technical), then layer content, structured data, and the discoverability files agents expect. Each row is one rule.

Foundation

The technical baseline. Get these wrong and nothing else matters.

  • glossary: Comprehensive SEO glossary; every term defined and grouped by category. Start here if any term is unfamiliar.
  • technical: Crawlability, canonicals, sitemaps, URL hygiene, rendering choices.
  • content: Topical authority, intent matching, internal linking.
  • e-e-a-t: Experience, Expertise, Authoritativeness, Trustworthiness.
  • helpful-content-update: Google’s helpful-content guidance, now part of core ranking: the site-wide signal and the pre-publish checks.
  • ai-content-policies: What Google’s and Bing’s rules allow and forbid for AI-assisted publishing, plus the compliance audit.
  • algorithm-history: Timeline of major named Google ranking updates, Panda through the helpful content system.
  • core-updates: Broad core and spam updates since 2018: 2025 to 2026 rollout dates, recovery guidance, and shared themes.
  • audit-checklist: Master single-page audit framework: pre-launch and quarterly.
  • on-page-audit-checklist: On-page, content quality, and internal linking audit tables: headings, titles, originality, AI-draft review, link depth.

AI search and agent readability

How to be quotable by AI engines and discoverable by autonomous agents.

  • llm-seo-best-practices: Pillar playbook for LLM SEO: topical clusters, literal-query titles, answer-first openings, structured metadata, and llms.txt.
  • generative-engine-optimization: GEO, AEO, and AI search optimization: how Google AI Overviews, Bing Copilot, ChatGPT Search, and Perplexity source answers, which crawlers to allow, and how to measure citations.
  • ai-citation-tracking: Measure AI citations per engine: Bing AI Performance report, fixed prompt sets, manual AI Overviews checks, and server logs.
  • answer-first-content: Lead every page with a direct, self-contained answer so LLMs can extract and cite the opening chunk.
  • structured-data-for-ai-crawlers: Which JSON-LD schemas help AI crawlers parse and cite a page, and why unbacked markup earns nothing.
  • ai-overviews: Earn inclusion and a citation link in the Google AI Overviews SERP feature.
  • llm-discoverability-standard: The six core items every site ships to be legible to AI assistants; the standard this domain anchors.
  • identity-page: The /identity brand page: fonts, type scale, color palette, logo, favicon, and media assets with a YAML token block.
  • llms-txt: The /llms.txt standard for agent-facing site indexes.
  • markdown-twins: Serve each page’s raw markdown at the URL plus .md and advertise it in llms.txt and the page head.
  • agents-txt: The /agents.txt agent-policy file: required keys, semantics, and validation.
  • ai-txt: The ai.txt standard for AI training opt-out.
  • discoverability-files: Master catalog of every special file a public site ships at the root.
  • site-metadata-files: security.txt (RFC 9116), humans.txt, web app manifest, favicon, Apple touch icon, default OG image.

Site architecture

URL hierarchy, redirects, scale.

  • site-architecture: Silo structure, URL hierarchy, breadcrumbs, depth-from-home.
  • internal-linking: Anchor text, link depth, orphan audits, three-click rule.
  • redirects: 301 vs 302 vs canonical, chain depth, query handling.
  • redirect-errors: Redirect chains, loops, noindex and 404 targets, soft 404s, homepage dumps, and the curl audit.
  • hreflang: Multi-language and multi-region site setup.
  • international-url-structure: ccTLD vs subdomain vs subdirectory for locales, and what hreflang does and does not fix.
  • pagination-and-facets: Large-site pagination and faceted-nav SEO.
  • crawl-budget: Crawl budget for sites over 10k URLs.
  • indexnow: IndexNow protocol for sub-hour change notification.
  • sitemaps-deep: Sitemap index, per-format sitemaps, news and image sitemaps, lastmod hygiene.
  • sitemap-extensions: Format-specific sitemaps: one child file per format, the 48-hour news window, image entries under the parent page.

Off-page and verticals

Link earning, and the vertical content types that need their own playbook.

  • backlinks: Backlinks and link building: referring domains, original research, broken-link rebuilding, source citations.
  • link-spam: Link spam and disavow: paid links, rel attributes, anchor-text mix, and when the disavow tool is warranted.
  • video-seo: VideoObject schema, transcripts, hosting, video sitemaps.
  • news-seo: Time-sensitive ranking, NewsArticle schema, Publisher Center, AMP retirement.
  • ecommerce-seo: Product schema, OOS handling, faceted nav, review schema.
  • ecommerce-product-urls: Status codes for out-of-stock and discontinued products; one canonical URL for color and size variants.

Content strategy

What to write, how to cluster it, how to keep it fresh.

  • content-clusters: Pillar pages plus cluster pages and the rules that bind them.
  • keyword-research: Tools, intent classification, long-tail terms.
  • title-tags: Title tag SEO as its own discipline: length, brand placement, click-magnetism.
  • serp-features: How to win featured snippets, People Also Ask, and knowledge panels.
  • featured-snippets: Paragraph, list, and table snippet structures, plus per-passage opt-out controls.
  • content-refresh: When and how to refresh existing pages without triggering HCS downweight.
  • content-pruning: Keep, merge, noindex, or delete underperforming pages; preserve URLs when consolidating.
  • cannibalization: Detect and fix two pages competing for the same query.
  • cannibalization-fixes: Merge with a 301, differentiate intent, or noindex; never canonical between editorial pages.
  • author-entities: Connect authors to Google’s knowledge graph to strengthen E-E-A-T signals.

Structured data

The markup that turns pages into rich results and AI citations.

  • structured-data: JSON-LD schemas worth shipping, and which still earn rich results.
  • schema-markup-deep: Product, Course, VideoObject, DefinedTerm, validation, and the visible-content rule.
  • retired-rich-results: HowTo, FAQ, sitelinks search box, and the June 2025 phase-out: what is gone, when, and what to do with the leftover markup.

Performance and rendering

Core Web Vitals, page speed, and the JS-side fixes.

  • core-web-vitals: LCP, INP, and CLS thresholds and fixes.
  • page-speed: TTFB, FCP, image optimization, font loading, JS deferral. What CWV does not cover.
  • render-blocking-resources: Critical CSS, font-display swap and font subsetting, and deferring or interaction-gating JavaScript.
  • javascript-seo: How crawlers render JS, render-budget reality, hydration cost.
  • spa-seo: SPA patterns that hide content or links from crawlers, and the URL Inspection and view-source checks.
  • mobile-first: Mobile-first indexing, viewport, touch targets, responsive.
  • mobile-usability: Viewport meta tag, touch target sizes, readable font sizes, and touch-action for tap delay.

Images and social

OG cards, image SEO, and the social-sharing surface.

  • og-images: OG image sizing and dynamic generation.
  • image-seo: Filenames, alt text, dimensions, formats, lazy loading, sitemaps.

Local

Geo-bound discovery.

Howtos (step-by-step)

When you need the procedure, not the rule.

Glossary anchors

Atomic definitions you can link to from any page.

65 items under this folder.