This week we’re using AI for two things that feel small but quietly change how your days go: **spring cleaning with a custom checklist** and **giving clear, kind feedback to a team member**. One helps you take back your house without burning a whole weekend. The other helps you say what needs to be said at work without losing sleep over the wording. Both are places where people tend to procrastinate, and both are surprisingly good fits for AI - if you tell it exactly what you need.

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

Use Case 1 of 2

Use Case 1: Spring cleaning with a checklist AI creates

On a Thursday evening in late May, Diane, 57, sat in her kitchen staring at a legal pad. She’d written “SPRING CLEANING” at the top and then stalled out. Every time she tried to list tasks, her brain offered up 50 things at once: clean out the pantry, wash curtains, deal with that mysterious box in the hall closet. She’d start in one room, find something that belonged in another room, carry it there, notice the dust on a bookshelf, and suddenly she was “cleaning” three rooms at once and finishing none.

She wasn’t lazy. She was overwhelmed. Her husband traveled for work, her knees weren’t what they used to be, and weekends disappeared into errands and caregiving for her mother. She knew a clean, reset house would make her feel calmer. She also knew that her usual approach - “do everything this Saturday” - ended in exhaustion and half‑finished jobs.

What she really needed was someone to break the job into sensible pieces, in a good order, that fit her actual life and energy. Someone who would think through the plan for her. That “someone” can be AI, if you ask it the right way.

Here is the exact prompt to use:

Copy and paste this into ChatGPT or Claude:

Copy and paste this into ChatGPT, Claude, or another AI assistant:

> I want you to act as a personal household project planner and create a realistic, room‑by‑room spring‑cleaning checklist for my home. > > Ask me 5 - 8 clarifying questions first about: > - how many rooms I have (kitchen, bathrooms, bedrooms, living areas, storage areas) > - whether I live in a house, condo, or apartment > - how much time I realistically have per day (in minutes) > - any physical limitations (e.g., bad knees, can’t lift heavy things) > - whether I will have help from other people > - my priorities (for example: kitchen and bathrooms must be spotless; I don’t care about the garage) > > After I answer, create: > 1) A 7‑day schedule with specific tasks for each day, grouped by room. > 2) A single master checklist I can print, with checkboxes and simple, concrete actions (e.g., “wipe baseboards in living room” rather than “deep clean living room”). > 3) A SHORT “start here” list of no more than 5 tasks I can do in 30 - 45 minutes if I get overwhelmed. > > Make sure the tasks are: > - realistic for my time and energy > - in a logical order (declutter before deep cleaning, top‑to‑bottom in each room) > - specific enough that I don’t have to think about what each one means. > > At the end, ask: “Would you like me to break any day into smaller 15‑minute chunks?” and, if I say yes, create that as well.

Why this prompt works:

You’re not just asking, “Make me a cleaning list.” You’re telling the AI what *role* to play (project planner), insisting it ask questions first, and giving it clear instructions on the format (7‑day plan, master checklist, short “start here” list). You also define what “good” looks like: realistic, logical order, concrete tasks. That forces the AI to think about your life, not some imaginary person with endless energy.

AI has no idea how fast *you* clean. It might give you an ambitious Day 3 that would take you four hours, not one. If a day’s plan looks impossible, tell it: “This is too much for me; reduce each day to 45 minutes and move the rest to an extra week,” and let it revise. Also, ignore any tasks that don’t fit your reality - if you don’t own a ceiling fan, you don’t need to dust one, no matter how authoritative the checklist sounds.

Use Case 2 of 2

Use Case 2: Giving feedback to a team member

On a Tuesday afternoon, around 3:30, Miguel, 49, stared at an email draft he’d rewritten four times. One of his analysts, Jenna, was missing details in her reports. Clients weren’t angry yet, but they were asking more follow‑up questions. Small things, but stacking up. Miguel needed to address it before it turned into a performance problem.

His problem wasn’t knowing *what* to say. He had very specific examples. His problem was saying it in a way that was honest without being harsh, and clear without sounding like a personal attack. In his twenties, he’d had a manager who gave vague “you need to step it up” feedback and then blindsided him at review time. He swore he would never do that to his own team.

So instead, he did what many conscientious managers do: he procrastinated. He told himself he’d “wait for the right moment,” which usually meant waiting until the situation was worse. What he needed was a first draft - a version of the message that was structured, clear, and respectful, in his own tone. Something he could edit, not invent from scratch.

That is a perfect job for AI, if you give it enough detail and boundaries.

Here is the exact prompt to use:

Copy and paste this into ChatGPT or Claude:

Copy and paste this into ChatGPT, Claude, or another AI assistant:

> I want you to act as a communication coach to help me give clear, kind feedback to a team member about their work. > > First, ask me these questions and wait for my answers before drafting anything: > - What is the team member’s role and how long have they been in it? > - What specifically is not working in their recent work (with 1 - 2 concrete examples)? > - What is working well that I genuinely appreciate? > - What outcome do I want from this feedback in the next 4 - 6 weeks? > - Do I plan to give this feedback by email, in a meeting, or both? > - What is my natural tone (for example: straightforward, diplomatic, warm, etc.)? > > After I answer, do the following: > 1) Draft a brief feedback message I can use as a script or email. > Use this structure: > - one sentence acknowledging something they are doing well > - 2 - 3 sentences describing the specific issues, in plain language, with examples > - 1 - 2 sentences explaining the impact on clients/the team > - 2 - 3 sentences being clear about what “better” looks like in the next 4 - 6 weeks > - one sentence expressing confidence in their ability to improve > > 2) Suggest 3 - 4 open‑ended questions I can ask them in conversation to invite their perspective (for example: “How does this land for you?” or “What would help you make this change?”). > > 3) Offer 2 - 3 alternative phrasings for the “tough” sentences so I can choose the one that sounds most like me. > > Keep the tone: > - respectful but honest > - specific, not vague > - focused on behaviors and outcomes, not personality. > > At the end, ask: “Would you like this to sound more direct, more gentle, or stay as is?” and be prepared to revise the wording based on my answer.

Why this prompt works:

You’re telling the AI to behave like a **coach**, not a script‑spewing robot. You force it to ask for context first, which means the feedback will be anchored in your real situation, not generic management clichés. The structure you give (what’s working, what’s not, impact, what “better” looks like, confidence) mirrors what good feedback training recommends, so it nudges the AI toward clarity instead of vague “be more professional” language.

AI loves to sound overly formal or overly cheerful. Before you use the draft, read it once and ask yourself: “If they only saw this message, would they clearly understand what needs to change and by when?” If the answer is no, tell the AI: “Make this 20% more direct while staying respectful,” and adjust until it matches your real intent. Also, never let AI invent details you didn’t provide; if it adds a compliment or criticism that isn’t true, delete it. Your credibility depends on accuracy, not polish.