One prompt. It builds the scorecard, then writes your day 45 check-in conversation word for word.
I keep saying I need help. Turn that into a role I could actually hire against, then write the conversation I will need to have with that person at day 45. Before you do any work, ask me these one at a time and wait for each answer: - Can you read my email, my calendar, and my task list? If I am a real estate agent, can you also read my text messages? - Which part of the business is drowning? Name it and I will confirm. - Where should the finished scorecard and conversation be saved? Then look at where my hours actually go before you ask me anything else. The work I would hand over is already visible there. Do not ask me to describe my week when you can read it. Name any connection you cannot reach. Show me what you found, then ask me one question at a time until you can answer all of these. Do not ask them all at once, and wait for each answer: - What would this person do in their first week? - What would be true in 90 days if they were great? - What would be true in 90 days if they were a disaster? - Which part of this job is judgment, and which part is just doing? - What have I tried already that did not work? PART 1 - THE SCORECARD THE MISSION - one sentence. What this role exists to produce. THE OUTCOMES - three to five. Each measurable, each with a date. THE FIRST 90 DAYS - what week 1, day 30 and day 90 look like. THE EARLY WARNING - how I will know by day 45 that this is not working, stated as something observable, not a feeling. Then, before I go hire anybody, tell me which of the outcomes should not be a person at all, and why. Point at the repeating work you found in my email, calendar, and task list rather than reasoning from the job description. PART 2 - THE DAY 45 CONVERSATION Ask me when this person starts, or when I expect them to. Work out the day 45 date and check whether it is already on my calendar. If it is not, tell me the date and offer to draft the event. Do not create it until I say yes. Write the check-in conversation in three versions. Use the early warning from the scorecard as the test in all three. VERSION A - IT IS GOING WELL. What I say so that good work gets named specifically instead of praised vaguely, and what I ask so the next 45 days are harder than the last. VERSION B - IT IS UNCLEAR. The hardest one. What I ask to find out whether this is a person problem, a role problem, or a me problem. Write the actual questions, in order, and tell me what each answer would mean. VERSION C - IT IS NOT WORKING. What I say in the first thirty seconds so nobody leaves the room confused about where they stand. Direct, not cruel. Include what happens next and by when. For all three: - Write my lines out in full. I will not improvise this well. - Mark the one sentence in each version I will be tempted to soften, and tell me what softening it would cost. Then tell me what to watch for in weeks 2 and 3 that shows which version is coming, so I am not guessing on day 45. Draft only. Never send, book, or change anything without my yes. Finish with the one outcome I am most likely to keep doing myself even after I hire someone, and what that would cost me.
Recommended: Claude Code, Cowork, or Codex. These prompts work best in an AI agent that can reach your tools, rather than in a chat window. This one reads your email, your calendar, and your task list, and your text messages too if you are a real estate agent, and it drafts only. It never sends or books anything without your yes.