This week, we are looking at two ordinary situations where AI can be surprisingly useful: turning spring cleaning into a manageable plan, and giving a team member feedback that is clear, fair, and useful. Neither requires technical expertise. You simply need to describe the situation honestly and tell AI what kind of help you want.

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

Use Case 1 of 2

Use Case 1: Build a Spring-Cleaning Checklist

At 8:15 on Saturday morning, April Martinez stood in her kitchen holding a roll of rubbish bags, a bottle of glass cleaner, and absolutely no idea where to begin. She had promised herself she would “finally get the house sorted” before her daughter visited the following weekend. But spring cleaning was not one job. It was windows, cupboards, bedding, the garage, paperwork, winter clothes, dusty light fittings, and a dozen small tasks she could remember only after she had sat down.

April did not need motivation. She needed a sensible order of operations. More importantly, she needed a plan that matched real life: a three-bedroom house, one bad knee, and only two or three hours available each Saturday. An AI-generated checklist could turn a vague burden into a series of decisions she could actually make.

Here is the exact prompt to use:

Copy and paste this into ChatGPT or Claude:

> I want you to create a realistic spring-cleaning checklist for my home. > > My situation: I live in a three-bedroom, two-bathroom house with a kitchen, living room, dining room, hallway, small home office, and single-car garage. I live alone, have a mildly painful knee, and can comfortably clean for no more than three hours at a time. I want to spread the work over four Saturdays rather than trying to do everything in one weekend. > > Please create a practical plan with: > > 1. A short list of supplies to gather before I begin. > 2. A room-by-room checklist. > 3. The best order for doing the work so I do not create extra mess. > 4. A four-week schedule, with no more than three hours of work per Saturday. > 5. Tasks I can skip without feeling guilty if I run out of time. > 6. Suggestions for protecting my knee, such as choosing tasks that can be done standing or sitting. > 7. A separate list of jobs that may require another person or a professional, such as moving heavy furniture or working at height. > > Use plain language. Do not assume I have specialist cleaning equipment. Estimate how long each task will take, and group tasks into 20- to 45-minute blocks. Ask me up to five questions first if any missing information would significantly change the plan.

Why this prompt works: It gives AI the information that changes the answer: the size of the house, the available time, the physical limitation, and the desired timeframe. It also defines the format you want - supplies, sequence, schedule, time estimates, and priorities - rather than simply saying, “Give me a spring-cleaning list.” The instruction to identify tasks that can be skipped is particularly useful. A good plan is not one that contains everything; it is one that helps you decide what matters most.

Use Case 2 of 2

Use Case 2: Give Constructive Feedback to a Team Member

At 4:40 on Tuesday afternoon, David Chen opened a draft report from a colleague named Marcus. The work was thoughtful and contained several valuable ideas, but it arrived two days late and was difficult to follow. Important conclusions were buried in long paragraphs, and Marcus had not included two figures the client had specifically requested. David needed to discuss the problem before Friday’s meeting.

He did not want to send a blunt email in frustration. Marcus was capable, conscientious, and new to the team. At the same time, David knew that vague praise followed by “please tighten this up” would not help. He needed feedback that separated the quality of the work from the missed expectations, described the impact, and made the next step clear. AI can help organize that message - but only if you provide the facts and retain responsibility for the judgment.

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 invent facts, diagnose their personality, or make assumptions about their intentions. > > The team member is Marcus, who has been in the role for five months. I am giving feedback about a client report submitted this week. > > What he did well: > - He identified three useful patterns in the customer data. > - His research was thorough. > - He responded quickly when I asked him to check a source. > > What needs improvement: > - The report was submitted two days after the agreed deadline. > - Two figures requested by the client were missing. > - The main conclusions were difficult to find because they appeared in long paragraphs rather than a short summary at the beginning. > > Impact: > - I had to spend several hours reorganizing the report before the client meeting. > - We may appear less prepared than we actually are. > - The client will not have the two figures they requested unless we add them before Friday. > > Please produce: > 1. A short opening that acknowledges the strengths of the work. > 2. A clear explanation of the three issues, using observable facts rather than labels such as “careless” or “unreliable.” > 3. A description of the impact without exaggeration. > 4. Three practical expectations for the next report. > 5. Two supportive questions I can ask Marcus so I understand what got in the way. > 6. A concise version I could say in a private conversation lasting about five minutes. > > Keep the tone direct, respectful, and appropriate for a manager speaking to a valued team member. Do not make the feedback sound artificial or overly corporate.

Why this prompt works: It distinguishes **observations** from **interpretations**. “Submitted two days late” is observable; “does not care about deadlines” is a judgment. It also explains the impact, which helps the recipient understand why the issue matters, and asks for specific future expectations rather than general criticism. Finally, the prompt tells AI what not to do: invent motives, use personality labels, or exaggerate. Those boundaries make the resulting language more useful and safer.