LLM (Large Language Model)

Definition

Large language model: an AI model trained on text that reads and writes natural language. In revenue tools it is the layer that turns numbers into sentences — morning summaries, report narratives, chat answers — and turns your questions back into data queries.

What it tells you

The LLM layer changes how you access your data, not what the data says. Its strength is language: explaining a pickup shift in two readable paragraphs, answering “how did leisure do last month?” without you building a report. Its weakness is that fluent language is not evidence of accuracy — the numbers behind a well-written sentence still need a source.

How to track it

Not a metric to track, but a component to recognise: when a tool summarises, narrates or chats, that’s the LLM; when it forecasts or recommends a rate, that’s typically machine learning on your historical data. Knowing which layer produced an output tells you how to verify it.

Where it fits

One half of the machine layer in modern RM — the language half, alongside machine-learning pattern recognition. The free Academy covers the division of labour in depth: AI and RM — what the machine sees and what it doesn’t.