Search used to be ten blue links. Now an answer engine reads a handful of pages and speaks for them. Being one of those pages is the whole game, and it is winnable because almost nobody is trying yet.
What you'll need
- Access to your site's head or page settings.
- Your most important three or four pages in mind.
Get your pages AI-ready
- Add schema.org markup. Mark up your key pages with structured data: Organization, Product, FAQ, Article. It tells machines exactly what your page is and means, instead of making them guess.
- Write a sharp title and meta description. These are the first things any model reads. Say what the page is and who it is for, plainly, in the first few words. No cleverness, just clarity.
- Add an llms.txt file. Drop a plain text file at yoursite.com/llms.txt that lists your most important pages and what they cover. It is a simple map for language models, the way robots.txt is for crawlers.
- Answer the real questions. Publish FAQ-style content that answers, word for word, what people ask about your space. Models quote pages that already say the thing the user asked.
Do it with AI
Generate the structure with Claude or ChatGPT, then paste it in.
Here is one of my key pages: [paste the URL and a short description]. Generate valid schema.org JSON-LD for it (pick the right type: Organization, Product, FAQ, or Article). Return the script tag ready to paste into the head, and tell me in one line where to put it.
My site is [domain] and it does [one line]. My most important pages are [list a few with one-line descriptions]. Write me an llms.txt file in the standard format that points AI models at these pages and explains what each covers.
Validate the schema with Google's Rich Results Test before you ship it.
How to check it worked
Ask ChatGPT or Perplexity a question your page answers, and see if your site shows up or gets cited. Then watch your referrals over the next few weeks for ChatGPT, Perplexity and Gemini appearing as sources.
Common mistakes
- Thin pages with no structure. Models skip pages they cannot parse. No schema and a vague title means you are invisible to them.
- Writing for keywords, not questions. AI engines answer questions. Pages phrased as real questions and answers get quoted far more often.
FAQ
- What is GEO?
- Generative Engine Optimisation, the practice of getting your content surfaced and cited by AI answer engines like ChatGPT, Gemini and Perplexity. It is to AI search what SEO is to Google.
- What is an llms.txt file?
- A simple text file at the root of your site that tells language models which pages matter and what they cover. Think of it as a friendly map for AI, similar to a sitemap or robots.txt.
- Does schema markup really help with AI engines?
- Yes. Structured data removes the guesswork, letting models understand exactly what your page is. Clear, marked-up pages are far more likely to be read and cited.
See AI traffic the moment it starts.
Holy Bucket recognises ChatGPT, Gemini, Perplexity and Claude as their own channel, so you know the day an answer engine starts sending you people.
It is a small channel for most founders today, and a growing one. We help you spot it early and lean in.