If you have read anything about AI prompting before, you will have come across frameworks that include two more components: role and context. And they are worth understanding, because there are situations where you will want to use them.
Role is the instruction that tells the tool what perspective or expertise to adopt. For most of your day-to-day work, your context document is already doing this: the tool knows you are a luxury travel advisor with a specific client base and a specific market position. You do not need to repeat that in every prompt.
But there are tasks where you want the tool to adopt a different lens on top of your baseline. If you are writing content aimed at a first-time traveller audience, you might add: “For this piece, write as if the reader has never been abroad before and needs everything explained without condescension.” If you are stress-testing a proposal, you might add: “Approach this as a sceptical client who is comparing three advisors and has not yet decided.” That kind of situational role instruction, layered on top of the permanent context document, is a powerful tool. Use it when the specific task requires a perspective that your context document does not cover.
Context, similarly, is not about who you are: the context document handles that. It is about what is specific and relevant to this task that the tool cannot know from your standing brief. The history of this particular client relationship. The fact that this proposal is a second attempt after a previous one was declined. The constraint that this client’s partner is not yet fully committed to the trip and the email needs to do some persuasive work without being obvious about it. Situational context that is specific to this moment, this client, this task. When you have it, include it. It will almost always improve the output.
