Sometimes you just need to say what you want, why you want it, and what a good result looks like. That's it. For example: "Can you draft a proposal section for our cybersecurity bid? We're trying to expand into federal cybersecurity work, so focus on our compliance and past performance. I'd like it to clearly show our technical edge."

If you've done something similar before, tell the AI about it. Give the situation, what you need done, what result you're after, and an example of when it worked. Something like: "We're bidding on a secure communications contract for a defense agency. Can you help draft a response to the RFP? On a past project, we built a similar system for a state agency and saw a 20% efficiency gain. We want to bring that same track record into this proposal."

The more moving parts a task has, the more helpful it is to spell out the task, exactly what you're asking for, the background, and an example. This works well for things like technical proposal sections where compliance and specs really matter.

When the output format matters most

Sometimes what matters most isn't the content, it's how it's packaged. In that case, tell the AI what role to take on, what the task is, and exactly what format you want back (a report, a table, bullet points, etc.).

For anything with real complexity, like assembling a full proposal, it helps to lay out the AI's role, your objective, the situation, what a good outcome looks like, and the steps to get there. This is the most detailed framework, but it's worth it for high-stakes or many-moving-parts work. Here's what that looks like in practice:

"You are a proposal manager coordinating a bid submission. Your objective is to win a contract to provide cloud-based services to a government agency. The RFP requires detailed technical specifications, a track record with sensitive data, and competitive pricing. Assemble a team across technical, financial, and compliance to build the proposal, addressing every RFP requirement while highlighting our unique value. Steps: review the RFP, assemble the team, draft each section, review and edit, then finalize for submission."

Explain the situation, name the problem, describe the action you want taken, and say what result you're after. Handy for revising a proposal that didn't land well the first time.

Lay out the situation, any complications or risks you're up against, your objective, your plan to get there, and how you'll measure success. This one's built for going up against tough competition.

Give the context, point to past examples worth modeling, define the AI's role, set the tone you want, give specific instructions, and specify the format. This is the most thorough option, best saved for final, client-ready documents.

Feel free to mix and match any of these, and don't hesitate to reach out to Customer Success if you'd like help building the right prompt for what you're working on.