How to get your website cited by ChatGPT and Perplexity: 7 steps backed by research

What the research says

The best evidence so far is GEO: Generative Engine Optimization by Aggarwal and colleagues, accepted to KDD 2024 (arXiv:2311.09735). The authors built a large benchmark of real-world queries (GEO-bench) and tested how edits to a page change how much of it an AI answer uses. The top methods, measured by position-adjusted word count against a baseline of 19.5 (results table, v3):

The abstract sums it up as a boost of "up to 40%". The same study warns that results vary by topic, so test on your own pages.

Step 1: let the right crawlers in

AI answer engines use separate bots for search and for training, and robots.txt treats them separately:

• OAI-SearchBot decides whether your pages can appear in ChatGPT search. GPTBot is for model training, and blocking it does not remove you from search (OpenAI's crawler docs).
• PerplexityBot surfaces and links sites in Perplexity answers and is not used for training (Perplexity's crawler docs).

Open yoursite.com/robots.txt. If either search bot is disallowed, AI search can't quote you, however good the page is.

Step 2: answer first

Put a 40–60 word direct answer at the top, before any story. Answer engines lift short, self-contained passages. If your first paragraph is a warm-up, the quotable part is buried.

Step 3: add numbers, with their source

Statistics were the second-best tactic in the study. "Most shops" is weak; a specific figure with a link to the official page that states it is quotable and checkable. Every number should link to where it came from.

Step 4: quote people who know

Quotations were the single strongest tactic. A line from a named expert, regulator or customer, with attribution, gives the engine something it can repeat with confidence.

Step 5: cite your sources inline

Link the study, the official page or the dataset right next to the claim. In the study this alone added about 28%.

Step 6: write the questions people ask as headings

Use the real question as an H2 or H3 ("How much does X cost?") and answer it in the next two or three sentences. Add an FAQ block with FAQPage structured data.

Step 7: don't chase a magic file

You may have heard you need an llms.txt file. It is a useful proposal from Jeremy Howard (September 2024) for giving AI tools a map of your site (llmstxt.org), and it costs little to add. But Google says plainly that you don't need new machine-readable files, AI text files or special schema to appear in its AI features (Google Search Central). Steps 1–6 matter more.

Check your page in 30 seconds

Our free SEO + AI-readiness score runs these checks on any public page, including which AI crawlers your robots.txt lets in. No sign-up.

Every source above was checked on 8 November 2026.