What Is LLM Visibility? How to Measure and Track It

What is LLM visibility?

LLM visibility is how often large language model assistants such as ChatGPT, Perplexity, Claude, Gemini and Grok name your brand when people ask them questions in your category, and how prominently they place it. It's also called AI visibility or AI search visibility.

It's measured on answers, not pages. Ask an assistant “best project management tool for a small agency” and it replies with a handful of names. Your LLM visibility is the share of such answers that include you, where you appear in them, and how that compares with your competitors. For ChatGPT specifically, see our guide to checking your brand visibility in ChatGPT.

Why LLM visibility matters

A search results page shows ten links and lets the buyer choose. An AI answer usually names three to ten products and explains why. Buyers who start there often go straight to one of those names, so the products left out of the answer lose the buyer before a comparison even starts.

It also compounds. Products that get named get clicked, written about and discussed, which puts them on more of the pages that assistants read next time. Getting into the answer early, while your category is still settling, is easier than displacing a product that every roundup already lists.

How LLMs decide which brands to name

An assistant can produce a recommendation in two ways, and most products need to win both over time.

1. From training data

The model has absorbed a huge amount of web text up to its training cutoff. If your product was widely and consistently described before then, the model may name you without searching. This changes slowly: only when a new model is trained.

2. From a live web search

For questions about the best or latest tools, assistants usually search the web, read the top pages and summarise them. Perplexity searches on every question; ChatGPT, Claude, Gemini and Grok search when the question needs current information. Here, your visibility depends on whether the pages that rank for the question mention you, and this can change within days.

In our own checks, one Medium article was enough for Claude to name a brand-new Mac app in an answer about window managers, while the established picks came from roundups that had ranked for years.

LLM visibility vs SEO

You'll also see the terms GEO (generative engine optimisation) and AEO (answer engine optimisation). They describe the work of improving LLM visibility; LLM visibility is the outcome you measure.

LLM visibility metrics

On DodoLaunch the visibility score gives an answer full credit for first place, 10% less for each place lower, and zero when you're not named, averaged across answers. You can try it on the AI visibility checker page.

How to measure LLM visibility credibly

Measure LLM visibility on a fixed set of buyer prompts, asked in clean sessions to several assistants, with each prompt repeated, and report trends rather than single answers.

• Clean sessions. Memory and past chats bias answers. Use temporary chats, logged-out sessions or the API.
• Enough prompts. 5 to 20 per topic, covering “best X”, “X for [buyer]”, “alternatives to [leader]”, price-led and problem-led questions.
• Repeat runs. Ask each prompt more than once; report a range when answers disagree.
• Several assistants. In one test we ran, ChatGPT, Perplexity, Claude, Gemini and Grok each gave a different list for “best window manager apps for mac”; only two apps appeared in all five.
• Same prompts every time. Changing the prompt set breaks the trend line.
• Record the sources. When an assistant searched, note which pages it cited. That's your to-do list.

How to run an LLM visibility audit

1. List your prompts. Write the 10 to 20 questions buyers ask before choosing a product like yours.
2. Run them across assistants. At least ChatGPT and Perplexity; ideally Claude, Gemini and Grok too.
3. Score each answer: named or not, position, competitors named, sources cited, and whether you're described accurately.
4. Find the gaps. Prompts where competitors appear and you don't are your priorities.
5. Trace the sources. Open the pages cited for those prompts. Are you on them? Could you be?
6. Check your own site. Does your homepage name your category in plain words? Is the text server-rendered? Does robots.txt allow AI search crawlers such as OAI-SearchBot and PerplexityBot?
7. Write the fix list in order of impact, and set a date to re-run the same prompts.

What to do with LLM visibility data

• Lost prompts: work on getting onto the pages the assistants cite for them.
• Wrong descriptions: fix them at the source (your site, your listings, old articles), then re-check.
• One engine much weaker: look at what that engine cites; each leans on different sources.
• Strong competitor: study the pages that name them; that's the coverage you need.
• Report trends: show mention rate and position over time, next to the changes you made.

How to improve LLM visibility

1. Be on the pages assistants read: roundups, comparison pages, launch platforms, reviews and community threads that rank for your category.
2. Use one clear description everywhere, with the category word buyers use.
3. Make your own pages easy to read: server-rendered text, a heading that names the category, short answers to buyer questions, structured data.
4. Don't block AI search crawlers in robots.txt.
5. Publish citable content: honest comparisons, original data, clear definitions.