ChatGPT, Claude and Perplexity are now where buyers ask "who's the best [thing] in [city]." If your business doesn't show up, the conversation is over before you knew it started.
The good news: getting cited isn't a black box. It comes down to a handful of technical signals AI models actually use to decide who's worth recommending. Most are fixable in an afternoon.
Here are the five we see brands miss most often โ in order of how fast they move the needle.
1. Unblock the AI crawlers in your robots.txt
Most sites have a robots.txt written for Google, not for AI. The result: GPTBot, ClaudeBot, PerplexityBot and Google-Extended get blocked by accident โ usually because the site uses User-agent: * with broad disallows, or because someone copy-pasted a "block AI" snippet from a 2023 blog post.
If a model can't crawl you, it can't cite you. Full stop.
Fix it now: add explicit Allow rules for the major AI bots at the top of robots.txt:
User-agent: GPTBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: Google-Extended
Allow: /
Not sure what's currently blocked? Run our free AI Crawler Checker โ it tests your robots.txt against 12 AI bots in one click and tells you exactly which ones to add.
2. Publish a llms.txt file
llms.txt is to AI models what sitemap.xml is to Google. It sits at the root of your domain and tells language models, in plain English, what your business does, what your key pages are, and how you want to be described.
It's not a ranking factor in the traditional sense โ but it's a citation factor. When models build their answer, having a clean, structured llms.txt makes you dramatically easier to summarise correctly. We've seen brands jump from "occasionally mentioned" to "default recommendation" after adding one.
Fix it now: generate yours in 30 seconds with our free llms.txt Generator. It produces a properly formatted file you can drop straight into your root directory.
3. Add Organization and FAQ schema
AI models love structured data. Not because they parse JSON-LD directly the way Google does โ but because schema gives them an unambiguous, machine-readable description of what your business is, where it operates, and what you sell. When two competing pages exist and only one has clean schema, the one with schema wins the citation.
The two that move the needle:
- Organization schema โ your name, logo, sameAs links to your social profiles, and contact info
- FAQ schema โ Q&A pairs that match how real users phrase queries to ChatGPT ("how much does X cost", "what's the difference between Y and Z")
Fix it now: if you're a VisibAI customer, the audit generates ready-to-deploy schema as a downloadable file. Otherwise, validate what you've got at schema.org's official validator before pushing live.
4. Make sure you're cited on the sources AI actually trusts
Here's the part most people miss: ChatGPT doesn't recommend brands based on your website alone. It recommends based on how often you appear on the sources it trusts โ Reddit, Wikipedia, YouTube, G2, Capterra, LinkedIn, and high-authority industry publications.
If you've never been mentioned on Reddit in your category, you're functionally invisible to AI for that category. This is the single biggest gap between brands that get cited and brands that don't.
Fix it now: spend 20 minutes searching Reddit for your category ("best CRM for agencies", "AI visibility tools", whatever). If your competitors are mentioned and you're not, that's your top priority for the next 30 days. Earn one good Reddit thread, one G2 listing, one Wikipedia mention โ in that order.
5. Restructure your content to answer questions, not rank for keywords
The old SEO playbook was: pick a keyword, build a page around it, optimise for density. The AI playbook is different. Models build answers by stitching together passages from across the web that cleanly answer a specific question.
If your homepage says "Leading provider of innovative solutions for modern enterprises", no AI is ever quoting you. If it says "VisibAI runs 52 queries across ChatGPT, Perplexity, Claude and Gemini to score how often AI recommends your business out of 100" โ that's quotable.
Fix it now: rewrite your hero, your About page, and your top three product pages so each opens with a single, declarative sentence that directly answers "what is this and who is it for". One sentence. No marketing fluff. AI models will lift that exact sentence into their answers.
The 30-minute version
If you do nothing else this week:
- Run the AI Crawler Checker โ fix any blocked bots
- Run the llms.txt Generator โ upload to your root
- Add Organization schema to your homepage
- Find one Reddit thread in your category and contribute genuinely
- Rewrite your hero into a single declarative sentence
That's the 80/20. Most brands won't do it, which is exactly why the ones that do start showing up in AI answers within weeks.
Want to know your actual score?
VisibAI runs 52+ queries across ChatGPT, Perplexity, Claude, Gemini and Mistral and tells you exactly where you're cited, where competitors beat you, and what to fix in priority order. The free audit covers ChatGPT โ no credit card, no signup friction.