Build the mental model

Think of a suitcase with a fixed volume: packing instructions, history, evidence, and the answer draft compete for the same space. Use the analogy to name the moving parts for context window, then drop it when you need numbers. For the contract-review assistant, the enduring idea is not a vendor feature name; it is the decision context window changes and the evidence that decision leaves behind.

The context window is shared working memory for one request—not permanent memory. What you pack, drop, or summarize changes outcomes.

Beginners often blur neighboring ideas when discussing context window. Keep it distinct by asking what artifact would still exist if model weights were frozen and only this layer changed on the contract-review assistant. If you cannot name that artifact, you are still describing “the AI” in general.

Tokenize inputs; if over limit, truncate/summarize/retrieve selectively; the model attends only to what remains; output tokens also consume the budget. In day-to-day language for context window: someone brings a need, the system inspects allowed evidence, this layer contributes a judgment or structure, and a consequence reaches a user or downstream system. Deterministic guards—permissions, schemas, arithmetic—still belong to the application around the contract-review assistant.

Uncertainty is normal for context window. Incomplete inputs and probabilistic behavior mean the contract-review assistant needs an escape hatch (retry, fallback, escalate) rather than fake certainty in fluent prose.

When you explain context window to a new teammate on the contract-review assistant, forbid the sentence “the AI just knows.” Replace it with the artifact that moves and the evidence you would file for context window. If they can falsify your picture with a single counterexample from last week’s traffic on the contract-review assistant, your mental model is working.

Also pin one numeric memory from this context window chapter: A 32k window with 8k system+tools, 12k retrieved exhibits, 6k history leaves ~6k for the draft—and less if tokenization is heavier than word count suggests. That number is not decoration; it is a template for how claims about context window on the contract-review assistant should look in design docs. Scoped specifically to context window / contract-review assistant / mental-model.

People confuse context window with neighboring buzzwords when debugging the contract-review assistant. Before changing prompts, ask whether the broken stage was evidence gathering, the context window judgment itself, validation, or the product action. Fixing the wrong stage creates folklore (“we tried context window and it failed”) that blocks the next team on the contract-review assistant. Scoped specifically to context window / contract-review assistant / mental-model.