Module 1 of 6
The context gap
Where clever prompting stops working.
Prompt technique gets you a long way and then hits a wall. The wall is always the same: the model does not know your prices, your policies, your product names, the way your team writes, or which of your three overlapping process documents is the current one. No amount of rephrasing fixes a knowledge problem, and most of the time people spend "improving prompts" is spent on exactly that mistake.
Run the comparison yourself before you build anything. Take a real recurring task, do it with a well-written prompt and no supporting material, then do it again with the right two documents pasted in. Note where the answers diverge. That gap — not a theory about how models work — is the specification for everything you build in this course.
When a task fails, ask whether the model lacked instruction or lacked knowledge. Prompting only fixes the first.