For twenty years, the homepage was the front door. A buyer heard your name, typed it in, and formed an opinion in eight seconds. That sequence is breaking. Increasingly the first impression happens inside an answer engine — ChatGPT, Perplexity, Gemini, or an AI overview stitched together from sources you never chose.
The uncomfortable part: you are being summarised whether or not you optimised for it. If your positioning is vague, the model will invent something plausible. If your category language is inconsistent across your site, reviews, and PR, the model will pick the loudest version — usually a competitor's.
AI optimisation is not a new channel. It is the discipline of making your expertise machine-legible: structured content, clear entity definitions, consistent naming, source-worthy data, and third-party corroboration. The brands winning citations are the ones publishing things worth quoting.
Key takeaways
- Define your category in one sentence and repeat it verbatim everywhere.
- Publish original data — models cite numbers, not adjectives.
- Use schema markup so machines can parse who you are and what you do.
- Track AI citations alongside rankings; they are a separate scoreboard.