By late September, “spring cleaning” may sound out of season - but the underlying problem is always timely: there is more to tidy than time to do it. And at work, giving feedback to a team member can feel just as difficult. This week, we will use AI for two practical jobs: creating a realistic home-cleaning checklist and preparing for a clear, respectful feedback conversation.

Try at least one before Sunday. That is the whole assignment.

Use Case 1 of 2

Use Case 1: Create a Spring-Cleaning Checklist You Will Actually Finish

At 9:15 on Saturday morning, Linda Morales stood in her kitchen holding a damp cloth in one hand and a scrap of paper in the other. She had promised herself she would “finally get the house under control” that weekend, but the promise had already become a dozen disconnected tasks: clean the refrigerator, sort the hall closet, wash the windows, dust the bookshelves, find somewhere for the winter blankets, and deal with the garage. Linda had about six hours across Saturday and Sunday. Her knees bothered her if she knelt for too long, and she was working alone. The problem was not that she lacked motivation. It was that she had no sensible way to decide what mattered most.

Here is the exact prompt to use:

Copy and paste this into ChatGPT or Claude:

> Act as a practical household project planner. Create a realistic spring-cleaning plan for my home, but do not assume I can clean all day or lift heavy items. > > Here is my situation: > - I live in a [house, apartment, or condo]. > - The rooms and areas I want to address are: [list them]. > - I have [number] total hours available, spread across [days and times]. > - I am working [alone/with another person/with family]. > - My physical limitations or safety concerns are: [describe them, or write “none”]. > - The areas causing me the most frustration are: [list two or three]. > - My main goal is: [for example, “make the home feel calmer,” “prepare for guests,” or “clear out clutter”]. > > First, briefly restate what you understand about my situation and identify any assumptions you are making. If important information is missing, ask me no more than five questions and wait for my answers. > > After I answer, create: > > 1. A day-by-day cleaning plan that fits within my available time. > 2. A room-by-room checklist using specific actions, such as “wipe the top of the refrigerator” rather than “clean the kitchen.” > 3. An estimated time for each task. > 4. A priority label for every task: “must do,” “helpful,” or “can wait.” > 5. A short list of tasks that can be skipped without undermining the overall result. > 6. A list of supplies I need, separated into “already likely at home” and “buy only if necessary.” > 7. A five-task, 30-minute version for a day when I lose energy or fall behind. > > Keep the plan practical, not perfectionist. Include breaks, avoid assigning heavy lifting unless I have help, and warn me if any task may require special equipment or professional assistance. Do not recommend mixing cleaning chemicals. Format the final answer with clear headings and checkboxes.

Why this prompt works: It gives the AI the information a generic request leaves out: available time, rooms, help, physical limitations, and the actual goal. It also asks for questions before recommendations, which reduces guesswork. The requested labels - must do, helpful, and can wait - force the plan to make decisions rather than produce an endless catalog of chores. Finally, the 30-minute fallback keeps one difficult day from turning into an abandoned project.

Use Case 2 of 2

Use Case 2: Give Clear Feedback to a Team Member

At 4:40 on Tuesday afternoon, David Chen opened an email from his colleague, Marcus, and immediately felt tired. Marcus was talented, well-liked, and usually dependable. But for the third time that month, he had delivered a project update late, and the figures in the latest version had not been checked against the source file. David needed to address the pattern before it affected a client deadline. He also knew that if he began the conversation with, “You need to be more reliable,” Marcus would hear a judgment about his character rather than useful information about his work. David had fifteen minutes before their one-to-one meeting and wanted help turning his frustration into a fair, specific conversation.

Here is the exact prompt to use:

Copy and paste this into ChatGPT or Claude:

> Act as a communication coach helping me prepare for a feedback conversation with a team member. Help me be direct, respectful, and specific. Do not diagnose the person’s attitude, personality, motivation, or intentions. Focus on observable behavior, impact, expectations, and next steps. > > Here is the situation: > - The team member’s role is: [role]. > - Their strengths and contributions are: [be specific]. > - The behavior or work issue I need to address is: [describe it]. > - The most recent example happened on [date or approximate date]: [describe exactly what happened]. > - Another relevant example is: [describe it, if applicable]. > - The effect on the team, customer, deadline, quality, or workload was: [describe the impact]. > - The standard or expectation is: [describe what should happen instead]. > - What I have already said or tried is: [describe it, or write “nothing yet”]. > - The support or resources I can reasonably offer are: [describe them]. > - The outcome I want from the conversation is: [describe it]. > > First, identify any wording that sounds vague, judgmental, exaggerated, or unsupported by evidence. Ask me up to five questions if you need more facts, and wait for my answers before drafting. > > Then create: > > 1. A 300- to 400-word conversation script in a calm, professional, human voice. > 2. A brief opening that acknowledges the person’s strengths without using praise as a way to soften or hide the issue. > 3. A clear description of the situation, observable behavior, and practical impact. > 4. Two or three open questions that invite the team member’s perspective. > 5. Two or three concrete changes we could agree to try. > 6. A simple follow-up plan, including when we will review progress. > 7. Three alternative phrasings for the most sensitive sentence. > > Use plain language. Avoid corporate clichés, labels such as “careless” or “uncommitted,” threats, and absolute words such as “always” and “never.” Clearly separate what I know from what I am assuming.

Why this prompt works: Good feedback depends on facts, not accumulated irritation. This prompt asks for the situation, behavior, impact, and expected standard, giving the AI enough context to help without inventing motives. It also instructs the AI to challenge vague or judgmental wording. Asking for questions, possible support, and a follow-up date turns feedback from a one-time complaint into a conversation about improvement.