A year ago a good title, clean subheadings, structured data, and solid links were enough to rank well in Google. In 2026 that is still the foundation, but no longer the whole game. Google AI Overviews, Perplexity, and ChatGPT with search answer the user without a click. To stay visible we need to be cited.
What content citability means
Citability is the ability of a single passage to stand on its own and answer a concrete question. Generative models rarely cite a whole article. They usually pick one well-formed paragraph, list, or table and build an answer around it.
How to write to be cited
- Answer one question in one paragraph, in the first sentence
- Use concrete numbers, dates, and sources instead of vague statements
- Use lists and tables because they are easier for a model to extract
- Add clear information about the author, publication date, and expertise
- Keep an unambiguous, factual tone and avoid marketing fluff
Structured data still matters
Schema.org is no longer only a hint for classic Google. LLM crawlers, including GPTBot, ClaudeBot, and PerplexityBot, use structured data to understand the context of the page. Article, FAQPage, and HowTo schemas have become particularly important.
You are no longer optimising for position one. You are optimising for the sentence the model will repeat back to the user.
Off-site brand signals
LLMs do not learn only from your site. Brand mentions on Wikipedia, Reddit, industry media, and academic publications increase the chance of being cited. It pays to keep NAP consistent, that is company name, address, and phone, and to maintain a healthy Google Business profile.
What to measure
Classic metrics like clicks and positions are still needed, but AI-answer visibility metrics matter more every month: citation count in AI Overviews, share of voice in Perplexity, and mention frequency in ChatGPT.