What is an LLM (large language model)?

A large language model, LLM for short, is an AI system trained on enormous amounts of text to understand language and produce it itself. Based on patterns, it predicts which word logically follows, and that is how a fluent, coherent answer emerges. LLMs are the engine behind ChatGPT, Gemini and Claude.

How does an LLM work?

During training, an LLM learns the statistical patterns in language: which words often appear together and in what context. It can then generate an answer to a question that fits those patterns. The model does not understand the way a human does, but it is remarkably good at recognizing structure, tone and facts it has seen often. This is a form of generative AI, because it creates new text instead of looking up existing text.

Where does an LLM get its answers from?

Partly from the text it was trained on, and in modern engines also from live-retrieved sources the model pulls in when a question comes up. As a result, a brand that is mentioned often and clearly can show up in answers more frequently. That is precisely why Generative Engine Optimization exists.

What does this mean for your business?

Because LLMs determine which brands appear in AI answers, it becomes important that your content is easy to understand and to cite. At Customer Impact we translate that into concrete visibility in AI engines that leads to enquiries. We work exclusively for B2B, with a small and fast team, and look at what a mention delivers in customers and revenue, not at vanity metrics.