A context window is the amount of text an AI can hold in mind at once — your conversation so far, anything you’ve pasted in, and its own replies. Go past that limit and the earliest material starts dropping off, and the AI can forget what was said or shared at the start.
Every AI assistant has a working memory limit called the context window, usually measured in tokens (roughly word-sized chunks of text). It covers everything currently "in view" for the AI: your messages, any documents you've pasted in, and its own previous replies in that conversation.
Think of it like a desk. You can spread out a certain amount of paperwork before there's no more room — and if you push more onto the desk, the oldest sheets fall off the back. That's exactly what happens to a conversation once it grows past the context window: the earliest exchanges quietly disappear from the AI's view, even though the chat history still displays on your screen.
This matters most when you're working with long documents or long conversations. If you paste in a lengthy contract early in a chat and then keep talking for a while, the AI may eventually lose track of the contract's details — not because it's being careless, but because that material has scrolled out of its working memory.
ExamplePaste one long report into a chat and the AI can discuss all of it accurately. Paste a dozen long documents into the same conversation, and by document ten it may have lost track of details from document one — not from confusion, but because the context window filled up.
If you're working with something long or important, keep the conversation focused, or start a fresh chat with just the document you need. Newer, larger models can hold much more before this becomes an issue, but every model has a limit.
See also
Learn to use this on your own work
Our AI Training & Enablement program takes non-technical Australian teams from the terminology to real, working AI builds on the job they already do.
Book a call →