The Automated Household

Build Your Own

From one 60-second prompt to a system that runs on rails.

Everything in the Stable System started with one prompt typed by one tired person at 5:30pm. You don’t need the architecture on day one. You need one win tonight, a rhythm this week, and a system this month. Here’s the actual path.

First: decide what your life is for

Skip this step and every automation you build will be an efficient way to do things that don’t matter. Before you automate anything, get explicit — uncomfortably explicit — about what you actually want:

  • What do I wish I was doing more of?
  • What do I value?
  • Where am I wasting time — not spending it on the people and activities I care about?
  • How do I maximize time doing what I want to do?
  • What do I want to get good at?

When I did this honestly, it took a week and felt like an existential meltdown. Years of flexing around everyone else’s needs meant I couldn’t name what I wanted anymore. What surfaced was one word: aliveness. Petroglyph hunting, watch restoration, building things, my people. And on the other side of the ledger: meal planning and grocery logistics, which made me feel precisely nothing. Those became the first candidates for real automation.

Automate what drains you. Protect what makes you feel alive. That’s the entire philosophy — the rest is mechanics.

Step one: the emergency meal prompt

The 5:30pm problem, solved in 60 seconds. Open Claude, ChatGPT, or Gemini. Take pictures of your pantry, fridge, and freezer. Type this:

The Emergency Meal PromptPlease review all of these images. These are the ingredients I have. My family of [X] is hungry and I have 30 minutes and no energy. Give me 3 dinner options I can make right now that will make an 11-year-old just as happy as a growing 17-year-old teen. Also I need some leftovers for lunch tomorrow.

That’s it. Three real options in under 30 seconds. This is the whole method in miniature: give the AI your actual context (photos, constraints, real humans with real preferences) and ask for decisions, not information.

Step two: find your rhythm

A meal plan is a decision made 5 times a week. A meal rhythm is a decision made once:

MonTueWedThuFriSat
Chicken + RiceMexican NightSoup / Stew / ChiliPastaSandwiches / WrapsTraeger / Grill

The rhythm is a framework your AI can work inside — shopping sales, prioritizing what your family actually eats, needing almost no correction. This one table cut my mental load in half.

The Weekly Meal Planning PromptPlan 5 dinners for our family of [X] this week. We like [cuisines]. Budget is $[X]. This week [store] has [items] on sale — build around those deals. At the end, give me the full shopping list. Also include any dietary restrictions, pet food needs, and household products we're likely running low on.

Once it’s dialed in, ask your AI to create a skill (Claude) or Gem (Gemini) or saved instructions from it — so you never explain it again. That’s the moment a conversation becomes a system.

Step three to five: the ladder

  1. From prompt to inventory

    A grocery list is just the output. The real system is inventory + meal planning + sale matching + purchase feedback. Photograph everything once, have AI build a working inventory, and feed it receipts and delivery confirmations. It stops asking you what you have — it knows.

  2. From inventory to operations

    Your household is an organization; AI is your operations partner. Chores, routines, kids' schedules across every school — the same pattern works: context in, rhythm defined, decisions out, everyone contributing on schedule via shared calendars and reminders. A chore chart is not the system. AI + reminders + proof + allowance + calendar working together is the system.

  3. From operations to rails

    Turn the whole thing into a repeatable, autonomous process: named agents with one job each, a rhythm keeper, and an audit loop that reviews what wasted and what worked. This is where the Stable System lives — and where you should feel free to get weird with it.

Make it yours — seriously

Here is the step everyone skips, and it's the one that determines whether your system survives: make it fun for you.

I named my agents after 1980s My Little Ponies and built them a console where they live in tree stumps in Fern Gully. That's not a confession — it's the reason the system still runs. Boring systems get abandoned. Systems you love get maintained, extended, and shown to friends.

Maybe yours are spaceships. Chess pieces. Characters from your favorite novel. With AI, the visualization, the personalities, the whole world costs you nothing but imagination — you just have to ask for it. Creativity isn't the garnish on this work. It's the fuel.

A bold pop-art painted horse mid-gallop — make the system unmistakably yours

Three things to do this week

Tonight: Think of one thing you want to do more of — and what you want AI to do in your household. Write it down in one sentence.

This weekend: Plan next week’s meals with AI. One conversation. Build your weekly meal rhythm and save it.

This month: Build one routine using a couple of agents — start small, let AI handle the heavy lifting, and protect the time it gives back for the thing you actually wanted to do more of.

Small steps. Smart systems. Big relief. You don’t need more time — you need a better system. And you’ve got this.

Take the field guides with you.

The workshop deck, the use-case finder, the system builder worksheet — all free.

Free Downloads