Glossary · Artificial intelligence

What is an LLM

An LLM (large language model) is software trained on huge volumes of text to predict and generate language. It's the engine behind assistants like ChatGPT and Gemini and behind AI agents. It's good at writing, summarizing and classifying, but it can make up information when it doesn't have the answer.

Almost every AI tool a company uses today has an LLM inside. Understanding what it is helps you know where to trust it and where not to.

What an LLM does well, and where it gets things wrong

In a sales operation, an LLM pays off on language tasks:

  • Good at: summarizing conversations, drafting messages and proposals, classifying requests, extracting data from text
  • Bad at: stating facts that aren't in what it was given, doing math without checking, deciding without rules

LLM, agent and assistant

The LLM is the engine. A chat assistant is the LLM talking to a person. An AI agent is the LLM with a goal, tools and rules, acting in other systems such as the CRM.

Frequently asked questions

Are an LLM and ChatGPT the same thing?

No. ChatGPT is a product; the LLM is the language model it uses under the hood.

Why do LLMs make up answers?

Because they generate the most likely text; they don't look up a source of truth. That's why, in sales use, they work with a defined knowledge base and rules about what they can answer.

Want to use AI with rules, not improvisation?

Nexo's free assessment shows where AI pays off first.

Get a free assessment Talk to us on WhatsApp