LLM

An LLM, or large language model, is an AI model trained on vast amounts of text to understand and generate language. It is the engine behind assistants such as ChatGPT, Gemini, Claude and the language models Perplexity uses. When you ask an assistant like that a question, the model predicts step by step which words are most likely to follow, based on patterns it learned during training.

That sounds simple, but because of the scale of the training, an LLM can do a surprising amount: summarise texts, answer questions, write code or translate. It does have limitations, though. An LLM doesn't automatically know what has happened since its training, and it can give answers that sound convincing but are wrong.

Where does an LLM get its knowledge?

Broadly speaking, there are two sources:

• Training data: the texts the model was trained on, such as public web pages, books and other sources. That knowledge has a cut-off date.
• Up-to-date sources when answering: many AI search engines search the web at the moment the question is asked and pass the pages they find to the model. That principle is called RAG, retrieval-augmented generation.

With ChatGPT's search feature, Perplexity or Google AI Overviews, you can see that second mechanism at work: the answer often contains references to sources. A source reference like that is called a citation.

Why do LLMs matter for your business?

More and more people ask questions directly to an AI assistant instead of searching on Google. Think of "which roofer in Lancashire specialises in slate roofs?" or "how much does an accountant cost for a sole trader?". The LLM then puts together an answer, sometimes naming businesses or sources. Whether your business shows up in it determines a new kind of visibility: AI visibility.

How do you increase the chance that an LLM mentions you?

There's no button to force a mention, and nobody outside the AI companies knows exactly how the models choose their sources. What does follow logically from how they work:

1. Make sure your website can be indexed by ordinary search engines and that you don't unintentionally block the crawlers of AI services.
2. Answer concrete customer questions clearly and factually, with a clear structure.
3. State explicitly who you are, what you do and which region you operate in.
4. Make sure your details are consistent across your website, your Google Business Profile and other listings.
5. Earn mentions on trustworthy external sites, such as trade media and industry associations.

That approach is the core of GEO, generative engine optimisation.

A realistic example

An online shop selling British craft beers notices that ChatGPT mainly names big international players when asked about gift hampers. By publishing clear guides, for example on the differences between an IPA and a pale ale, and by being mentioned on UK beer blogs, the shop gives both search engines and AI systems more useful material. Whether that leads to mentions can't be guaranteed, but you can measure it. On the platform that tracks your visibility in AI assistants, you can see how often and in what context LLMs mention your business today.