TL;DR
Paste your household’s members, recurring tasks, and constraints into one prompt and a current model (Claude Opus 4.7, GPT-5.5, or Gemini 3.1 Pro) returns a weekly grid, a quarterly rotation, and a “house rules” doc in about 15 minutes. The single most important output is the fairness audit: total minutes per person per week. If anyone is more than 30% above the household mean, you have just found the source of past arguments. Treat the result as a proposal to discuss together, then move the final version into a shared Google Sheet, OurHome (free), or Tody so it actually gets followed.
The problem AI solves here
Chores pile up, two people remember to do them, and the rest forget. The fix is not more nagging. It is a written, visible rotation everyone agreed to, with clear ownership and a sane way to handle weeks when someone is sick or traveling.
The hard part is the math, not the cleaning. Matching N recurring tasks (each with its own frequency) to M people while respecting work shifts, child ages, chemical allergies, and “who hates which task” is a constraint-satisfaction problem. That is exactly the kind of combinatorics a language model handles well once you give it the constraints. It also drafts the house-rules document in language that does not sound preachy.
What AI cannot do is sense fairness the way the people living together can. So the schedule it produces is a starting proposal, not a verdict. The non-negotiable step is sitting down together and letting everyone trade or veto one task before anything gets printed.
Which AI tool to use
All three current flagship models are strong at this. The difference is in how the output lands.
| Tool (as of June 2026) | Model | Cost | Best for |
|---|---|---|---|
| Claude (web/app) | Opus 4.7 or Sonnet 4.6 | Free tier; Pro $20/mo | Rendering the schedule as a live, editable Artifact you can tweak in place |
| ChatGPT | GPT-5.5 | Free (with ads in US); Plus $20/mo | Iterating in Canvas; exporting to Google Sheets/Calendar via connectors |
| Gemini in Google Sheets | Gemini 3.1 Pro | Google AI Pro $19.99/mo | Building the table directly inside a shared spreadsheet your family already uses |
For a one-time schedule, the free tier of any of them is enough. Claude’s Artifacts are the standout for this task: ask for the schedule and it opens an interactive table in a side panel you can edit without re-prompting. Gemini’s advantage is that it builds the grid straight into a Google Sheet (Insert the generated table), so the whole household can open it on any phone.
A short comparison of the underlying tools lives in our ChatGPT vs Claude vs Gemini guide.
What to feed the AI
The quality of the schedule is capped by the quality of these five inputs. Spend ten minutes here.
- Members: name, age, weekly availability (work shifts, school, the evenings someone is reliably out)
- Recurring tasks with frequency: dishes daily, vacuum weekly, deep clean monthly, gutters twice a year
- Hard constraints: who physically cannot do what (heavy lifting, chemical allergies, can-or-cannot reach the top shelf)
- Strong preferences: who hates which task, and the often-overlooked half: who actually enjoys one
- Slack rules: how many “skip” tokens each person gets per month before a swap is required
The “who enjoys what” line is the single biggest fairness lever. A task one person dreads and another mildly likes should never rotate onto the person who dreads it.
Copy-ready prompt
Replace each [bracket] placeholder with your own details, then paste the whole block in.
You are a household operations designer. Build a fair weekly chore rotation.
Members: [names, ages, weekly availability]
Tasks (with frequency): [task list, e.g. dishes daily, vacuum weekly, deep clean monthly]
Hard constraints: [who cannot do what]
Preferences: [who hates / who enjoys which task, per person]
Slack: [skip tokens per person per month]
Output, in this order:
1. Weekly schedule as a markdown table: rows = days, columns = each person,
cells = the task plus estimated minutes.
2. Monthly/quarterly rotation: who handles the less-frequent tasks
(deep clean, fridge, bathrooms), rotated every quarter so no one is stuck.
3. Annual list: holidays, hosting, seasonal tasks (yard, gutters), spread across the year.
4. Fairness audit: total weekly minutes per person; flag anyone more than 30%
above or below the household mean and suggest a swap to fix it.
5. House rules: 5 friendly bullets covering how to skip, swap, and escalate.
6. One paragraph: exactly what happens when someone is sick or travels for a week.
Rules: keep cooking and dishes on the same person the same night;
put a small daily task on whoever is already in that room;
do not optimize minutes so hard that tasks get chopped into tiny fragments.Then follow up with one of these, depending on your tool:
- Claude: “Render the weekly schedule as an editable Artifact table.”
- ChatGPT: “Put this in Canvas, then format the weekly table for export to Google Sheets.”
- Gemini (in Sheets): “Insert the weekly table into this spreadsheet.”
Reading the fairness audit
The fairness audit is the part most people skip, and it is the only part that settles arguments with numbers instead of feelings. Here is a worked example for a couple plus one 12-year-old:
| Person | Weekly minutes | vs. mean |
|---|---|---|
| Adult A | 195 | +18% |
| Adult B | 140 | −15% |
| Kid (age 12) | 60 | n/a (age-scaled) |
Score the adults against each other, not against the child, whose load is scaled to age. A gap inside roughly ±15% is normal and not worth fighting over. A gap above 30% is the model telling you to move one weekly task. Re-prompt with “swap one task from Adult A to Adult B to close the gap” rather than asking for perfect equality, which only fragments tasks into ten-minute slivers nobody wants to track.
Make it stick
A rotation that lives in one person’s phone notes gets forgotten. Move the final, agreed version to a shared surface:
- A shared Google Sheet is the lowest-friction option and works with Gemini’s “insert table” step.
- OurHome is fully free, syncs across all devices, and adds a points/reward system that works well with kids.
- Tody is built around room-by-room cleaning; its FairShare and multi-device sync sit behind a Premium plan. Sweepy is a paid alternative at about $8.99/month after the trial.
Pin or print a copy too. The app is for tracking; the printout is what stops the “I didn’t know it was my turn” conversation.
Run it for 30 days before changing anything
Treat the first month as a measurement period, not a draft. Follow the schedule untouched and log what actually happened: who did each task, who used a skip token, what silently went undone. After 30 days, sit down and adjust based on the log, not on what was supposed to happen.
Two metrics tell you it is working: weekly minutes per person converge to within ±15%, and the genuinely hated tasks rotate at least once a quarter. If a task never gets done by anyone, the answer is usually to automate it (a robot vacuum) rather than reassign it.
Common mistakes
- No slack for sick days, so one missed week collapses the whole rotation
- Pairing a loud task (vacuum) with a quiet expectation (toddler napping)
- Asking the model to be perfectly fair on minutes, which fragments tasks into unmanageable slivers
- Splitting cooking and dishes across two people on the same evening
- Quietly assigning chores to a partner instead of agreeing on the schedule together
- Never updating the schedule when someone changes jobs or schools
FAQ
Which model should I use if I only have free accounts? Any of them works for a single schedule. Claude’s free tier (Sonnet 4.6) plus Artifacts gives you an editable table without paying. The US ChatGPT free tier now shows ads but still produces the grid.
How do I handle a kid who refuses? Give age-appropriate tasks, attach a visible reward (OurHome’s points system is built for this), and rotate so no child is stuck with a hated chore for more than four weeks. Ask the model to “scale the 12-year-old’s load to roughly half an adult’s minutes.”
What about uneven work hours or income? Adjust expected minutes proportionally and be transparent about why. Ask the model to run a “what if” scenario, for example “recompute assuming Adult B works two evening shifts.” It will rebalance the daily tasks around the availability gap.
Does this work for roommates, not just families? Yes, and the constraints conversation matters more than the document. Start by collecting everyone’s hard constraints and hated tasks out loud, then let the model draft from that. A shared Google Sheet beats a group chat for keeping roommates honest.
How often should I redo it? Once a quarter. Re-prompt with the last three months of skipped tasks and the household’s reaction, and ask for tweaks only: swap a hated task, raise the slack tokens, or automate something. A full rebuild every month just creates churn.
Related
Tie weekly chores to your bigger plan with weekly planning prompts, reduce the task volume itself using a decluttering plan, and reuse this framework for a move with the moving checklist workflow.
Tags: #Productivity #Workflow