Spring cleaning and giving better feedback are two jobs that look simple from a distance and become surprisingly tiring once you begin. In both cases, AI can help you turn a vague intention into a practical plan: one that tells you what to do next, and one that helps you say something difficult without making the other person defensive.

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

Use Case 1 of 2

Use Case 1: Spring cleaning using a checklist AI creates

At 9:15 on Saturday morning, Susan Miller stood in her kitchen holding a half-empty bottle of glass cleaner, a laundry basket, and a list she had started three different times. She wanted to freshen the house before her daughter and two grandchildren visited in April, but “spring cleaning” had grown into an intimidating cloud of jobs: windows, closets, dusty ceiling fans, the pantry, baseboards, the guest room, and the pile of things in the hall cupboard. Susan had a bad knee, worked best in 45-minute stretches, and did not want to spend an entire weekend moving furniture. The problem was not knowing that the house needed attention. The problem was deciding what mattered most, what could wait, and what she could realistically finish.

Here is the exact prompt to use:

Copy and paste this into ChatGPT or Claude:

> I want you to create a practical spring-cleaning checklist for my home. Do not give me a generic list. Ask me up to 10 questions first, one at a time, about the rooms in my home, what needs attention, how much time I have, my physical limitations, supplies I already own, and whether anyone can help me. > > After I answer, create a checklist organized by room and priority. Divide it into: > 1. Must do before guests arrive > 2. Helpful if I have time > 3. Optional deep-cleaning jobs > > Break the work into sessions of no more than 45 minutes, with a 10-minute rest between sessions. Put the easiest high-impact tasks first. Include a short list of supplies I need to buy, but do not recommend expensive products unless necessary. Avoid tasks that require ladders, heavy lifting, or kneeling for long periods unless you clearly label them as tasks for another person. > > For every task, include an estimated time, the room, the supplies needed, and a checkbox. Also include a suggested order: declutter first, then clean from high surfaces down to lower surfaces, and floors last. If a task could damage a surface or requires following the manufacturer’s instructions, say so. > > At the end, give me a realistic two-day plan and a shorter one-day emergency version. Use plain language, not professional cleaning terminology.

The detail about working from high surfaces down to floors is useful because dust tends to fall as you clean. A sensible checklist also groups tasks by room and begins with decluttering, while remembering often-forgotten areas such as light switches, door handles, baseboards, filters, and the tops of cabinets. More importantly, this prompt tells the AI about Susan’s actual life: her time limit, her knee, her supplies, and the difference between essential and optional work.

Good prompts usually contain five ingredients: the situation, the desired result, the person’s limits, the format, and the standard for success. “Give me a spring-cleaning list” leaves nearly everything unspecified. The prompt above tells the AI to ask questions before making recommendations, create short work sessions, estimate time, and flag safety concerns. It turns AI from a source of random suggestions into a planning assistant.

Use Case 2 of 2

Use Case 2: Giving feedback to a team member

At 4:40 on Tuesday afternoon, Martin Alvarez opened a draft email from his colleague Jenna and immediately felt his shoulders tighten. Jenna had missed two deadlines in the past month and had sent a client document containing several avoidable errors. Martin was frustrated, but he also knew Jenna had recently taken on a new account and was caring for her father. He needed to address the pattern before it affected the client relationship, yet he did not want to send an angry message or deliver a vague lecture about “being more careful.” His real challenge was separating what he had observed from the story he was telling himself about it.

Here is the exact prompt to use:

Copy and paste this into ChatGPT or Claude:

> Help me prepare respectful, specific feedback for a team member. Do not write an angry email and do not diagnose their attitude, motivation, or personality. First, ask me questions one at a time so we can separate facts from assumptions. Ask about: > > - The person’s role and responsibilities > - The specific behavior or work product I observed > - When and how often it happened > - The effect on the client, team, schedule, or quality > - What standard or agreement was not met > - What the person may need from me > - What improvement would look like > > After I answer, give me three things: > 1. A short script for a private conversation > 2. A follow-up email that summarizes the agreed next steps > 3. Three questions I can ask to understand what is causing the problem > > Use a calm, direct tone. Describe observable behavior rather than character. Use this structure: what I observed, why it matters, ask for their perspective, agree on a specific next step, and set a date to review progress. Do not use phrases such as “you always,” “you never,” “you need to care more,” or “this is unacceptable” unless the facts clearly justify them. If my account is incomplete or unfair, point that out before drafting anything. Keep the conversation script under 250 words and the follow-up email under 150 words.

This prompt works because it forces a useful pause before the wording begins. It asks for observable behavior, timing, impact, and an agreed standard. “Jenna is unreliable” is a judgment. “The client draft arrived at 3:20 p.m. on Wednesday, one day after the agreed deadline, and contained four incorrect figures” is information someone can discuss and improve.

It also asks for the other person’s perspective. That matters because missed deadlines can have several causes: unclear priorities, unrealistic workload, missing information, lack of training, or a personal problem the manager does not know about. The prompt creates room to investigate without excusing the problem. Finally, requiring a date for review turns feedback into a management conversation rather than a moment of emotional release.

Know someone who spends too long on things AI could do in two minutes? Forward this to three people who subscribe and I will send you my free AI Prompt Starter Pack: 20 ready-to-use prompts for everyday life.

By Dr. Rowan Hayes