How to Write Content That Ranks in AI Search
Writing for AI search (ChatGPT, Perplexity, Google AI Overviews) requires a different structure than writing for traditional Google. The unit of optimisation has shifted from the page to the passage — short, quotable, fully self-contained answers.
1. Lead with the answer. Open every section with a direct, declarative response in 1–2 sentences. AI engines lift these verbatim.
2. Use question-based H2 and H3 headings. Structure your page as a series of "What is X?" / "How does Y work?" sections. This matches the prompts users feed into AI engines.
3. Add structured data. FAQPage, Article, and HowTo schema dramatically increase citation likelihood. Nest Person and Organization schema for author authority.
4. Be specific. Numbers, dates, sources, and named entities outperform abstract claims. "40% of UK searches end without a click in 2026" beats "many searches don't result in clicks."
5. Cover the entity, not the keyword. Build comprehensive topical clusters. AI engines reward depth across an entity, not isolated keyword pages.
6. Cite reputable sources. Outbound links to authoritative sources (.gov, .edu, established trade publications) reinforce your own credibility signals.
7. Refresh regularly. Update dates, statistics, and examples quarterly. Stale content gets demoted in AI training cycles.
Burying the answer 4 paragraphs in Vague intros ("In today's fast-moving digital landscape…") Walls of text without subheadings Keyword stuffing Thin pages targeting single keywords
See our full GEO playbook or book a free GEO audit.