Issue 39: Smarter by Thursday
Smarter by Thursday
One practical AI win, every week. No jargon required.
Spring cleaning and giving feedback to a team member may seem unrelated, but both become easier when you stop asking AI for “help” and start giving it a clear job. In one case, AI can turn a vague ambition into a manageable household plan. In the other, it can help you turn a difficult conversation into specific, respectful, useful feedback. Neither requires technical expertise - only enough detail to describe the situation honestly.
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 8:15 a.m. on Saturday, April 4, Margaret stood in her kitchen with a mug of coffee, a full dishwasher, three grocery bags on the floor, and the sinking feeling that “spring cleaning” had somehow become a second job. She wanted to clean the house before her grandchildren visited in two weeks, but every online checklist seemed written for somebody with unlimited time, strong knees, and a storage room. Margaret’s home had two bedrooms, one bathroom, a basement laundry area, and a garage she had been avoiding since 2022. She did not need a list of 75 chores. She needed to know what mattered most, what could wait, and what could be completed in short sessions without exhausting herself.
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 two-bedroom, one-bathroom home. > - The main areas are the kitchen, living room, bedroom, bathroom, laundry area, and garage. > - I have approximately 30 minutes on weekdays and two hours on Saturday. > - I want the house looking noticeably cleaner within two weeks. > - My priorities are dust, floors, kitchen surfaces, bathroom surfaces, clutter, and the entryway. > - I have limited energy and do not want tasks that require standing for more than 30 minutes at a time. > - I have ordinary household supplies: dish soap, all-purpose cleaner, white vinegar, baking soda, microfiber cloths, a vacuum, and a mop. > > Create a two-week plan divided into 30-minute weekday sessions and two-hour Saturday sessions. Put the highest-impact tasks first. Include a short supply list, but do not recommend buying specialized products unless genuinely necessary. Separate cleaning, decluttering, and maintenance tasks. For each task, tell me exactly what “done” means. Include one optional task for each session in case I have extra energy, and mark any task that may require help from another person. > > Keep the tone practical and encouraging. Do not give me more than five main tasks in any one session. Before creating the plan, identify any important safety issues, such as mixing cleaning products or using a ladder.
Why this prompt works: It gives AI the information a generic request lacks: the size of the home, available time, priorities, energy limits, supplies, and deadline. It also asks for definitions of “done,” which prevents vague instructions such as “organize the kitchen.” The request to separate cleaning, decluttering, and maintenance makes the plan easier to follow, while optional tasks protect you from feeling that an imperfect day means failure.
Use Case 2 of 2
## Use Case 2: Giving feedback to a team member
At 3:40 p.m. on Tuesday, September 22, Daniel closed his office door before replying to an email from Priya, a capable team member whose work had recently become unreliable. Three project updates had arrived late, and two contained figures that had not been checked against the shared spreadsheet. Daniel was frustrated, but he also knew that saying “You need to be more careful” would not help. Priya had been with the organization for four years, handled difficult clients well, and was already carrying a heavy workload. Daniel wanted to address the problem without embarrassing her or turning one performance issue into a personal judgment. He needed help organizing his thoughts before their one-to-one meeting on Thursday.
Here is the exact prompt to use:
Copy and paste this into ChatGPT or Claude:
> Help me prepare for a constructive feedback conversation with a team member. Do not write an insulting, threatening, or overly corporate script. Help me be specific, fair, and respectful. > > The situation: > - The employee is named Priya. > - She is experienced, thoughtful with clients, and generally dependable. > - Over the past three weeks, three weekly project updates were submitted after the agreed deadline of 3:00 p.m. Tuesday. > - Two of those updates contained figures that did not match the shared financial spreadsheet. > - The errors were caught before being sent to the client, but correcting them delayed the team’s work. > - I have not yet asked Priya whether there is a workload, process, training, or personal issue contributing to this pattern. > - I want the next update submitted by Tuesday at 3:00 p.m. and checked against the spreadsheet before submission. > > Create the following: > 1. A brief opening that explains the purpose of the conversation. > 2. A neutral description of the observable behavior, using the dates and facts above rather than labels about Priya’s character. > 3. A sentence explaining the impact on the team and client work. > 4. Three open questions I can ask to understand what is causing the problem. > 5. A clear, reasonable expectation for the next two weeks. > 6. A way to offer support without taking responsibility for doing the work myself. > 7. A closing that confirms the next check-in. > > Keep the complete draft under 350 words. After the draft, explain which sentences are facts, which are requests, and which are questions. Also list any assumptions you made that I should verify before the conversation.
Why this prompt works: It gives AI a balanced picture instead of presenting Priya as “a problem employee.” Specific dates, deadlines, errors, impact, strengths, and unknown causes create the raw material for fair feedback. Asking for questions is especially important because feedback should be a conversation, not a speech. The final request - to identify assumptions - helps you notice where you may be filling gaps with guesses.