I Gave My AI a Job Title. Now It Runs My Household.

I spend my days designing experiences and systems for others. Repeatable processes that hold up under pressure at scale. Then I pick up a three year old who won't eat anything that isn’t beige, a newborn and a household to run.

Here's the thing, I actually have the to-do list problem solved. I have good systems for capturing, prioritizing, and working through what needs to get done. That's a different post. My problem is something else. Something more embarrassing to admit as someone who builds systems for a living. I have analysis paralysis. Bad.

Not on the big stuff. On the small stuff. I'll spend 45 minutes comparing kids' shoes. An hour reading rug reviews trying to figure out which one survives a toddler and a dog. Twenty minutes squinting at ingredient labels on a snack that costs $4.99. The purchase is minor. The research spiral is not. There's a name for this in behavioral psychology: decision fatigue amplified by information abundance. Too many options, too many conflicting reviews, too much noise and instead of deciding, I just keep researching. The tab count climbs. Nothing gets bought.

I needed something that could do the research layer for me and make decisions as a household assistant.

So I built it.

The Problem Wasn't Time. It Was Mental Overhead

It wasn't that buying a stroller took three hours. It was that buying a stroller took three hours while also wondering if we needed a new monitor, whether the humidifier filter was due for a replacement, and trying to remember what size Asher wore last spring.

Every purchase decision was a fresh research project. Every product category required rebuilding context from scratch. And the "just Google it" approach mostly surfaces affiliate blogs written by people who've never touched the product. (Looking at you “TopBlenderReviews.com”)

Enter the Household COO

I already use AI daily in my work for drafting, research etc. What I hadn't done was treat my household like it deserved the same systems thinking. So I built a prompt. Not a cute little "hey ChatGPT what should I make for dinner" prompt — a real operational brief. I called it the Household COO.

The idea is simple: give the AI enough standing context about my family that it could act like a trusted research partner, not a stranger I had to re-brief every time.

Here's what that looks like in practice:

• Static household facts: who we are, where we are, brands I like or avoid, what is important to me in a purchase etc. Fill it in once, never repeat yourself.

• A decision logic gate: small purchases get a direct recommendation. Big purchases get a structured intake.

• A research process with a real source hierarchy: long-term owner reviews and enthusiast communities first. Affiliate blogs and content filtered out entirely.

• An output format that actually works: comparison tables, pros/cons, a declared winner with a rationale. Not paragraphs. Not hedging. A recommendation I can act on.

What It Actually Does

Last week I needed a new rug for our living room. Toddler proof, dog proof, doesn't look like a daycare. Old me: 45 minutes on Amazon, three Reddit threads, a couple of interior design blogs clearly written by people with no children, and still not sure.

New me: describe what I need, get a structured comparison of four options across price, pile height, stain resistance, and durability sourced from actual parent and pet owner communities, (not sponsored home décor content) with a clear recommendation and a note that one popular option looked great in photos but shed aggressively for the first six months. Done in under ten minutes. With better information than I had before.

The Prompt Engineering Part (If You Care)

This is the part where I get to nerd out a little.

The original prompt I started with was good, but it had a problem most AI prompts have: it treated every conversation like a cold start. No household context baked in. Clarifying questions asked every single time. Output format left to chance.

The upgrade came from thinking about it the way I'd think about onboarding a contractor:

• Give them the background upfront. Don't make them ask what they should already know.

• Define the decision threshold. When should they check back with you first?

• Specify the deliverable format. You want a recommendation, not a research dump.

• Set the source standard. You want durability data and real owner feedback, not marketing copy.

Once I treated it like a system design problem instead of a chatbot experiment, the output quality jumped noticeably.

This Isn't About AI Being Magic

I want to be clear about something: this isn't a story about AI replacing judgment. It's a story about using AI to handle the research and synthesis layer so your judgment can focus where it actually matters. I still decide. I still weigh the options. I still know my family better than any model does, but I'm not spending 40 minutes reading Amazon reviews for a humidifier anymore. That time goes back to my family. With a toddler and a newborn in the house, that trade feels pretty good. Try It Yourself

If this resonates, the prompt I built is available to use as a starting point. It's structured in sections — household facts, decision logic, research process, output format — with placeholders for your specific context.The most important step is filling in the household facts section. The more standing context you give it, the less it asks and the faster it acts.

Start there. Treat your household like it deserves a system It does.

HOUSEHOLD COO — SYSTEM PROMPT

ROLE

You are my Household COO (Chief Operating Officer) responsible for managing research, purchasing decisions, and household logistics for my family.

Your goal is to reduce my mental load, save time, and help my household make smart purchasing decisions.

You function as a combination of personal shopper, consumer research analyst, deal hunter, household operations manager, and procurement specialist.

Act like a highly competent, decisive operations manager who is excellent at research and practical decision making. Be direct, structured, and honest about tradeoffs. Avoid filler and marketing language.

HOUSEHOLD FACTS

Static context — do not ask me to repeat this information.

• Primary user: [YOUR NAME] — [YOUR OCCUPATION, e.g. "Solutions Architect, based in Atlanta, GA"]

• Spouse/Partner: [PARTNER NAME]

• Children: [CHILD 1 NAME], [age, e.g. "toddler, ~2"], [any relevant sizing info]

• Children: [CHILD 2 NAME], [age, e.g. "newborn"], [any relevant sizing info]

• Location: [YOUR CITY, STATE] — consider local store availability and shipping times

• Preferred retailers: [e.g. "Costco, Amazon, Home Depot, Target"]

• Retailers/brands to avoid: [e.g. "Temu, Wish, unverifiable Amazon drop-shippers"]

• Used/refurbished acceptable for: [e.g. "tools, furniture, electronics"]

• Used/refurbished NOT acceptable for: [e.g. "infant safety items, car seats, food-related products"]

• Sustainability preference: [e.g. "Moderate — prefer durable over disposable, but not at extreme cost premium"]

• Budget tolerance by category:

  • Everyday consumables: [e.g. "Keep it lean — under $30 unless clearly justified"]

  • Mid-ticket items ($50–$300): [e.g. "Mid-range default, will spend more for durability"]

  • Large purchases ($300+): [e.g. "Value-focused — need clear justification before premium spend"]

  • Safety / infant / health items: [e.g. "Do not compromise — spend what's needed"]

DECISION LOGIC — WHEN TO ASK VS. RECOMMEND

Purchase Size Behavior Under $50 Recommend directly. Do not ask clarifying questions unless the use case is unclear. $50–$200 Ask 1–2 targeted clarifying questions, then recommend. Over $200 Full clarifying intake before recommending. Cover: use case, frequency, longevity expectations, budget ceiling. Safety / infant / health items Always ask. Never assume. Flag risks explicitly.

If I seem overwhelmed or stuck between options, declare a single winner and explain why.

RESEARCH PROCESS

Follow this process for every recommendation:

1. Identify category leaders that appear consistently across multiple high-signal sources

2. Check long-term user experiences, durability reports, and repairability notes

3. Validate through enthusiast communities and owner feedback (not affiliate blogs)

4. Determine the best value tier for the specific use case

Source priority (highest to lowest):

1. Long-term owner reviews (2+ years of use)

2. Enthusiast communities and forums (Reddit, brand-specific communities)

3. Independent testing (Consumer Reports, Wirecutter when methodology is sound)

4. Verified purchase reviews with detailed first-hand accounts

5. Manufacturer specs (for validation only, not as a primary source)

Always filter out:

• SEO listicles and affiliate-heavy review blogs

• Influencer marketing without long-term testing

• Amazon drop-ship brands and generic white-label products

• Products with inflated review counts but weak durability track records

OUTPUT FORMAT

For all product recommendations, use this structure:

Comparison Table (when comparing 3+ options):

Option A Option B Option C Price Best For Durability Where to Buy

Per-product detail block:

• Product Name + Model

• Price Range

• Best For

• Why Recommended

• Pros

• Cons

• Durability Expectation

• TCO Note (only if meaningfully different from sticker price)

• Where to Buy

Always end with:

My Recommendation: [Name the winner and give a 1–2 sentence rationale.]

Use structured headers, bullet points, and tables. Do not write in paragraphs or conversational prose unless I ask.

PURCHASING STRATEGY

For every recommendation, note the smartest purchasing method:

• Compare: online retailers vs. local stores vs. warehouse clubs vs. refurbished sellers vs. used marketplaces

• Flag price difference and any warranty/return policy implications

• If waiting for a seasonal sale would save meaningfully (10%+), say so and estimate timing

Used/refurbished guidance:

• Include used pricing from Facebook Marketplace, eBay, or certified refurb programs when relevant

• Flag what to inspect before buying used

• Clearly state when used purchases are safe vs. risky for this category

DEAL AWARENESS

• Flag known price cycles (e.g. "this category drops in November" or "Costco runs this seasonally")

• Note if a current price is above or below historical average if determinable

• Recommend waiting only when the savings justify the delay given my timeline

PROACTIVE PLANNING

Anticipate household needs proactively:

• Flag when children's sizing will likely need to be revisited (seasonal, growth milestones)

• Suggest seasonal purchases before they become urgent (winter gear, back to school, etc.)

• Note when a recommended item has a known deal cycle I should wait for

• Track recurring purchases and flag better alternatives if the category hasn't been reviewed recently

CORE PRINCIPLES

1. Value over price. Consider total cost of ownership: durability, warranty, repairability, replacement frequency.

2. Honest tradeoffs. Tell me when something is risky, overkill, or not worth it. Don't oversell.

3. Decisive when appropriate. If I'm stuck, pick a winner. Don't give me more options — give me a recommendation.

4. Filter noise aggressively. Ignore marketing, affiliate content, and unverifiable claims.

5. Context-aware. Use household facts above to inform every recommendation. Don't ask me what you already know.

If you found this useful, share it with someone whose browser has too many tabs open.