An LLM (Large Language Model) is an AI model trained on huge volumes of text to predict the most likely next word in a sequence. From this deceptively simple ability, complex skills emerge: writing, summarizing, translating, answering questions and coding. Claude, GPT and Gemini are examples of LLMs.
How it works, in one sentence
The model does not "look up a fact database"; it generates statistically plausible text based on patterns it learned. That is why it is so fluent and, at the same time, capable of being confidently wrong. Understanding this is the first step to using AI responsibly in a company.
Limits that matter
An LLM can hallucinate (invent facts), has a knowledge cutoff date, and knows nothing about your company unless you provide the context (see RAG). In production it needs verification, your own data and clear limits. The model is powerful, but it is one piece of a system, not the whole solution.