| Make a page eligible for Google AI Search surfaces | Ensure it is indexable, snippet-eligible, crawlable, and useful in Google Search | Google's generative AI features are rooted in Search ranking, retrieval, and quality systems |
| Optimize for GEO/AEO on Google | Treat it as SEO, not a separate hack track | Google's guidance frames AEO/GEO work as optimizing the Search experience |
| Implement technical foundations | Use crawlable links, canonical URLs, correct status codes, rendered primary content, sitemaps, and stable metadata | Google must discover, render, understand, and select the page before content work matters |
| Improve content for AI answers | Add unique experience, original facts, clear structure, media, and user-satisfying coverage | Non-commodity, people-first content is the durable signal |
| Use structured data | Add only page-representative JSON-LD that matches visible content and Google feature docs | Structured data helps eligibility for rich results but is not required for generative AI search |
| Control snippets carefully | Use nosnippet, max-snippet, max-image-preview, and data-nosnippet only when the tradeoff is intentional | Restrictive preview controls can also limit use in AI Overviews and AI Mode |
| Ignore unsupported hacks | Do not create llms.txt, AI-only markup, artificial chunking, fake mentions, or mass query-variation pages for Google | Google explicitly lists these as unnecessary or risky for Google Search |
| Support browser agents | Use semantic controls, labels, stable layouts, visible actions, and clean accessibility tree signals | Agentic experiences inspect screenshots, HTML, and accessibility trees |
