LLM SEO: getting ChatGPT, Gemini and Claude to mention you

What LLM SEO means

An LLM, or large language model, is the technology behind ChatGPT, Gemini, Claude, Perplexity and DeepSeek. When someone asks one of these assistants a question, the assistant writes an answer. That answer may name businesses, products or services. LLM SEO is the work of making those models more likely to mention your business, accurately, when people ask about what you offer.

You may also hear this called GEO, which stands for generative engine optimization, or AEO, which stands for answer engine optimization. They describe the same work. The goal is simple: when a customer asks an assistant for a recommendation in your field, you want to be in the answer, described correctly.

This matters because more people now start their search in an assistant instead of a search engine. For example, a dentist in Leeds might ask an assistant for a good family dentist accepting new patients. The assistant gives a short list. If you are not on it, you were never in the conversation.

Two ways an LLM learns about you

There are two paths, and they behave very differently. Knowing which one you are working on saves you time and stops you chasing the wrong thing.

Training data is what the model absorbed from the web before its cut-off date. This is slow to change. You cannot edit it, and you cannot pay to be added. If a model learned about your business a year ago, that version may still shape some answers.

Live search is different. Many assistants search the web before they answer, then cite the pages they found. This is where your changes show up sooner. If you publish a clear page today, an assistant that searches live can find it and use it in an answer this week.

For most small businesses, live search is where the practical work happens. You are not trying to rewrite the past. You are making sure the pages and mentions an assistant can find today say the right things.

LLM SEO checklist

This is the practical part. Work through these in order. None of them require technical skills, and each one makes it easier for an assistant to describe you correctly.

1. Let AI search crawlers read your site. Crawlers like OAI-SearchBot and PerplexityBot visit pages so assistants can use them. Check your site settings and allow them unless you have a specific reason not to. If you block them, you are invisible to live search.
2. State clearly on your site what you do, where, for whom and at what price. A visitor should understand this in the first few lines. For example, a plumber in Manchester might write: emergency plumbing in Manchester, homes and small offices, call-out from a fixed fee, available seven days.
3. Answer common questions in short, self-contained paragraphs. Each answer should make sense on its own, without the reader having seen the rest of the page. Assistants often lift a single paragraph, so do not bury the answer in a long story.
4. Be present on the review sites, directories and lists that assistants cite in your field. Find out which sources assistants quote when people ask about your type of business, then make sure your details are on those sources.
5. Keep details consistent everywhere. Your name, address, phone number, opening hours and services should match across your site, your profiles and any directory listing. Mixed details make an assistant less confident about you.
6. Correct wrong information at its source. If an old page or listing says you closed, moved or changed what you offer, fix that page. Publishing a new page does not remove an old mistake.

The order matters less than finishing the list. A business that does all six will be easier for an assistant to describe than one that only rewrites its homepage.

If you want to understand how this differs from traditional search work, see our guide to GEO vs SEO.

What an LLM SEO tool should do

You can do this by hand, but it is repetitive. You would need to ask the same questions every week, across several assistants, and remember what changed. A tool is mainly useful for the checking and the tracking.

• Ask several models your customers' questions on a schedule. One assistant is not enough, because they answer differently. You want to see the pattern across models.
• Show full answers, competitors named and sources cited. A summary is not enough. You need to read what was actually said, see who was named instead of you, and see which websites the assistant used.
• Turn gaps into specific actions. Knowing you are missing is only useful if it leads to a fix. The best output is a ranked list of things to change, not a score.
• Keep a record over time so you can tell whether your changes worked.

Vasukai does this from $19 a month, with a free first check that shows what ChatGPT says about your business. It asks the assistants your customers' questions on a schedule you pick, shows the full answers, who was named instead, and which sites were cited. It then gives a ranked list of actions, and a button that drafts the text for each fix. It also includes Google SEO tools such as keyword research, weekly rank tracking and a monthly technical site check.