Large language model (LLM) — a model trained on very large amounts of text to predict the next piece of text, which turns out to be enough to draft, summarize, classify, translate and answer questions.
Why it matters for a business
It is the engine inside ChatGPT, Claude, Gemini and most of what is now sold as “AI.” When an assistant answers a question about your category, an LLM wrote the sentence. When an automation drafts a reply to a new lead, an LLM wrote that too.
Two properties worth remembering
- It is fluent by design, not factual by design. It produces the most plausible continuation. Plausible and true overlap a great deal, but not completely — see hallucination.
- It knows what it was trained on, up to a cutoff. Anything newer, or anything specific to your business, has to be supplied at the moment of the question. That is what retrieval and grounding do.
“LLM” and “AI” get used interchangeably in most sales conversations. They are not the same thing: a lead-scoring model or an image classifier is AI and is not an LLM. It matters when someone is quoting you a price.
Related: retrieval-augmented generation, grounding, hallucination, AI agent.