AI at work
Context window and tokens
The context window is how much an AI can read and keep in mind at once in one conversation. It is measured in tokens, chunks of text roughly three quarters of a word long. Usage is often billed by the token.
Why it matters for your business
A bigger window means the assistant can read a long contract or a whole conversation without forgetting the start. It also means a higher cost per answer. For most business uses the limit is not a problem, but it explains why some tools "forget" and why usage bills vary.
In practice
A law firm's document summarizer handles a long lease in one pass because the model's context window fits it. A smaller model would need the file split into pieces.
Related terms
Want this handled rather than explained?
Two minute intake. A real person reads every one and replies within a business day.