Context window

Definition

The amount of information an AI model can work with at one time, including instructions, conversation history, documents and other supplied context.

Put simply

It is the AI's available working space for the task at hand.

What boards and leaders should understand

A large context window does not mean AI will use everything in it equally well.

Performance can deteriorate when conversations or agent workflows run too long, topics become mixed, instructions accumulate, or the system receives too much irrelevant, outdated or poor-quality information.

So when AI output deteriorates, the underlying model may not be the problem. The way context is being assembled and managed may be.

Questions to consider

  • Where we rely on long-running AI workflows, how are we managing the quality and relevance of the context they accumulate?