LLM (Large Language Model)

An LLM (Large Language Model) is an artificial intelligence model trained on enormous amounts of text to predict the next word in a sequence. From that apparently simple mechanism emerge capabilities such as answering questions, summarizing documents, translating, writing code and holding coherent conversations.

An LLM works with tokens (pieces of words) and responds to prompts (the instructions the user sends). It does not "browse the internet" by default: it answers based on what it learned during training, which explains limitations like the knowledge cutoff and hallucinations (plausible but incorrect answers). Techniques like RAG and fine-tuning exist precisely to mitigate this.

In practice, LLMs are consumed in two ways: through chat interfaces (ChatGPT, Gemini, Claude) or through APIs, which let you embed the model in your own products. A concrete example: a support team uses an LLM to automatically summarize every closed conversation and classify the contact reason, feeding reports with no manual work.