A large language model (LLM) is an AI model trained on vast amounts of text. It can understand language, summarise it and generate text of its own. LLMs power ChatGPT, Google Gemini and many other AI assistants.
How does an LLM work?
At its core, an LLM predicts which word is most likely to come next. From huge volumes of text, it has learned how terms, facts and relationships are typically connected. This results in two properties that matter for businesses:
- The knowledge from training has a cut-off date. The model only knows newer information if it searches the web live.
- An LLM can make statements that sound convincing but are wrong. These are known as hallucinations.
That is why AI search tools such as ChatGPT, Perplexity and AI Overviews combine the model with a web search and name the sources they find.
Why do LLMs matter?
More and more people ask an AI before choosing a medical practice, an online shop or a piece of software. What an LLM ”knows” about your brand depends on how often and how clearly it appears in training data and current sources.
For businesses, this means Google remains important, but search is getting broader. With ongoing GEO management, you make sure AI models know your brand, describe it correctly and mention it in relevant answers.
