Iterate With Specific Instructions, Not Vague Requests

Most reps treat AI like a vending machine: put in a question, take out an answer, move on. This approach produces mediocre results. The reps who get the best outputs treat AI like a conversation.

The iteration principle: Never accept the first output as final. Always follow up with at least one refinement prompt.

The most effective refinement prompts are specific about what to change:

• "Make it more direct. Cut the second paragraph."
• "The tone is too formal. Rewrite in a more conversational style."
• "The opening is too generic. Start instead with the specific trigger event: [TRIGGER]."
• "Make it 40% shorter without losing the core message."
• "Add more specificity to the ROI claim in paragraph 2."

What doesn't work: "Make it better." This gives the AI no direction and often produces a version that's different but not improved.

A useful workflow: Generate the first draft → identify the single weakest element → give one specific instruction to fix it → repeat once more if needed. Three prompts total, 5 minutes, output you can actually use.