I was in a “society,” which is the women’s college equivalent of a sorority, but decidedly less fun and certainly not as cool.

This is a fact I rarely think about anymore, except when I try to pose naturally in a photograph or casually throw a sweater around my shoulders without looking like a linebacker.

There is a particular strain of female competence that I associate, perhaps unfairly, with women who went to SEC schools and joined sororities.

They know where to put their hands in photos. They know how to make their hair work in 90% humidity.

I do not. And it turns out, I am not alone.

I was recently on a rare girls’ trip and we were all crammed in front of a mirror, getting ready to go out together (which is by the way my all-time favorite activity, a large group GRWM). And it started.

“What are you doing with that curling rod? That is not what I do. Is that how you get your curls to stay?!”
“Oh, wait. How do you actually blow-dry your hair with so much volume?”
“Is that contour? How are you using it?”
“Have you tried peptides?”

This is how it always is. Everyone has picked up a few things along the way, but nobody has the whole Alpha Kappa Mamma manual.

The prep session concluded with my friend blow-drying all of our hair. Not because we’d planned a salon. Because once one person admitted she didn’t know, everyone admitted it, and the one who could do it just… did it. Five heads and one round brush blow dryer. So much volume!

Here’s what I realized somewhere around head three: I have spent years assuming everyone else got the manual and I’d been out sick. I went to an all-girls high school and a women’s college. Eight years, fully immersed, in all things female. But here I was at 46, in a room full of accomplished women, and none of us had it. There is no manual. There was never a manual. There’s just people telling each other the next thing.

And it gets better, because whatever manual you did cobble together at some point, it changes. What worked at 30 does not work at 46. The face is different. The hair is different. The internet is full of women who will gladly sell you the new version, except most of it is noise and you don’t find that out until you’ve bought the noise. Someone rewrote the manual, but nobody sent the update.

What changed for me isn’t that I finally learned all of it. It’s that I stopped waiting to absorb it by osmosis and started just asking. I ask my AI Stylist an astonishing number of questions on a weekly basis. Very few are sophisticated, but all are cringe.

So many photos and quick conversations.

The blush.

My face is round and, 70% of the year, quite pale. I need blush, but it never seems to be as sculpted as I want, even with contour. So finally, I took a picture of my face and asked Chat to apply blush virtually and show me exactly how to place it. That’s how I finally figured out my own blush placement. One photo. A decade of looking faintly walloped, resolved.

How to stand in a photo.

This is the one I’m a little defensive about, so I’ll just say it. I do not know how to stand. In photos I look braced. In video, I look like a hostage being filmed for proof of life, slightly too still, eyes saying, “tell them I’m okay.” I have known this about myself for years and done nothing, because what is there to do, take a class?

I took a picture. I asked. Where do I put my weight, what do I do with the arm that’s just hanging there, why does everyone else look at ease and I look like I’m waiting to be processed. The fixes were almost insultingly small. Weight on the back foot, one thing to do with my hands so they have a job. I tried them in the mirror feeling ridiculous. Then I looked at the photo. So much better.

The waist that isn’t where I think it is.

Here’s a more useful one. I am 5’1”, but I have a naturally high waist. I did not know this in any actionable way. I just knew that everything I wore had a faintly Empire-dress quality, like I was perpetually one ribbon away from a Jane Austen adaptation, and I assumed that was just how clothes looked on me.

It is not just how clothes look on me. I sent a few photos and asked why everything read slightly off. The answer was about where I’d been putting belts and where my waist actually is. Two different places, it turns out. Moving one changed the entire silhouette. I had been styling a body I’d guessed at instead of the one I have.

The part.

Fast one. My hair part needed to move. Not a little, considerably farther over than felt natural. I’d been parting it where I’d always parted it, which is to say where someone established it for me at approximately age nine. I asked. I moved it. It was right. Twenty seconds, a thing I’d carried unexamined for thirty years.

The video call.

I know how to do my makeup. Twenty years of practice, a decade of doing other people’s before events. In a mirror, I’ve got it.

Then I started getting on podcasts and Zoom calls, and something was off. I’d look in the mirror, feel good, walk to my desk, turn on the camera, and watch myself go flat. Paler. Somehow more tired than I’d looked ninety seconds earlier. Same face. Same makeup. Completely different result.

So I took a photo of myself in my actual setup, at my actual desk, in my actual lighting, and asked Chat what needed to change for camera specifically.

The answer was not “more makeup.” I braced for “more makeup.” It was more definition, in very specific places. Slightly stronger brows. More blush than felt right. A little more contrast around the eyes. And permission to stop being afraid of looking overdone in person, because the camera eats all of it.

That was the part that got me. “Too much” in the bathroom mirror is “finally visible” on a webcam. The two are not the same room, and I’d been styling for the wrong one.

Every one of these was a small, specific, fixable thing nobody had told me.

There’s no manual. There never was. The SEC girls didn’t get one either, they just had better lighting and more reps.

Fewer things I’ve quietly assumed were just how I look. Fewer decades spent parting my hair where a stranger left it. Eight years of women’s education, and I finally learned to stand in a photo. From a chatbot. At 46.

That’s your little win.

In part because my AI operating system needs a reset for fall.

Yes, I realize it is the middle of August. But the start of school feels like fall, and I am as irritated by that as I am by pumpkin spice appearing in stores before Labor Day.

Last week, I gave AI the entire school calendar and asked it to find the friction before we were living it.

But this week, when the backpacks came out and the school emails started flying, I realized that not only am I am in summer mode, but so is my operating system.

My OS desperately needs a reset for fall.

The girls have new teachers. Fall sports are starting. The sports apps are waking back up. School emails are arriving again (and again, and again). Lunches need to be packed and dinner suddenly has to fit around soccer, choir, homework, etc.

The shape of our week has changed and I need my weekly brief to catch up quickly.

Here’s what goes into the reset:

1. What are we doing now?

Summer camps are over and sports are back. I need Claude to understand our normal weekly rythym so it can tell me what is on deck – and more importantly, notice when something changes.

I don’t need the Weekly Brief to simply repeat what is already sitting on my calendar. I can read a calendar. I want it to give me the lay of the land for the week and flag the things that require some action on my part: a practice that moved, a game that was added, something they need to bring, or a schedule change that makes Tuesday night suddenly much more complicated than it looked before.

2. What people and systems get special attention?

My inbox gets a seasonal adjustment as well. A few weeks ago, I cared about camp emails and vacation confirmations. Now I have two new teachers, school announcements, sports notifications, activity emails, and approximately 4,000 other messages that accompany the start of a new school year.

So I’m playing favorites and teaching my OS who matters the most by identifying key names, apps and email aliases:

- Teacher names

- School email alias

- App names and email aliases

I’m also updating the kids’ grades, so that AI surfaces a school announcement that affects my elementary school kids, but does not pay attention to an eighth grade event.

It all comes to the same inbox, but it doesn’t all need to crowd my headspace.

3. When is normal most likely to break?

The girls are walkers this year. They walk home from school. I already have a weather app, so I don’t need a full weather report in my Weekly Brief. I do need to know if Wednesday afternoon’s weather means the normal after school plan isn’t going to work. Thunderstorms at 3:00 matter in a way that rain at 10:00 p.m. doesn’t. If I’m going to need to rearrange my afternoon to pick them up, I’d rather know that on Sunday than realize it while I’m in the middle of a client call on Wednesday.

The same principle applies to food. As I’ve said before, I do not do elaborate meal planning, but I do want to understand which nights look particularly crunched. If everyone is home and we can actually sit down together, great, that’s a real dinner night. If we have soccer at 5:00 and choir at 6:30, that is not the night for culinary ambition. I need to know that before I find myself staring into the refrigerator at 5:47, or going through the Taco Bell drive thru at 6:35.

I’m doing something similar with school lunches. The girls order lunch on certain days – usually pizza on Fridays and occasionally other days – and we pack the rest. I don’t need a full AI-generated lunch plan. I just need a reminder that says, “pizza Friday, but you’re packing the other 4 days”.

Weather, lunches and dinner sound like three completely different things, but they’re really the same instruction: show me where our normal plan isn’t going to work.

That’s the difference between a Weekly Brief that summarizes my life and one that actually makes my life easier.

4. And finally, put it where I’ll actually see it.

I changed the timing and delivery mechanism for the Weekly Brief. I used to review it Sunday night – now it will be delivered Sunday morning, when I still have enough time to do something about the week ahead. If we need groceries for four packed lunches, a pickup plan because Wednesday looks like a washout, or a dinner solution for a night when everyone is going in different directions, I would much rather figure that out over Sunday morning coffee than Sunday night.

And I want it by text.

The point of all this isn’t to build the most sophisticated AI system. The point is to have the right information show up at the right moment, so I have fewer things rattling around in my head.

Last week’s school-calendar exercise gave me the big picture for the year. This week’s reset is much more mundane: teachers, soccer practices, packed lunches, rain.

It’s about making sure the system understands the life we’re actually living.

Because fall arrived at my house today, whether I like it or not.

This year, my stress level is exacerbated by the fact that our school did not release the calendar until the spring. At that point, most of us had already registered for a full summer of camps and scheduled August vacations when we found out the year would start a week earlier than it used to.

And not with a normal first week - there is a one-hour meet-and-greet in the middle of the workday on Wednesday, followed by two three-hour half days beginning Thursday.

My youngest is enrolled in a full week of camp that week. Of course, it is the most expensive camp of the summer with the longest hours (9:00 – 4:00). But instead of coordinating one relatively easy camp drop-off and pick-up, I am now trying to figure out whether she skips camp, misses part of school, or spends the week being shuttled between the two like a tiny executive with an impossible calendar.

I am annoyed about all of it.

But mostly, I am just not ready to go back to school, and neither are my kids.

I cannot change the school calendar. But I can stop being surprised by it.

So this year, I gave the entire calendar to AI and asked it to show me the friction before we were living it: half days, long weekends, childcare gaps, possible vacation windows, and the family dates we needed to protect.

Here’s what I got in return:

• A clean list of school closures, half days, and early dismissals in my preferred format

• A ranked list of the dates most likely to create childcare problems

• Suggested travel windows for trips we want to take, optimized to ensure that the kids miss as little school as possible

• A list of decisions and reminders to auto-calendar

I also asked AI to count the number of half days, early dismissals, and school closures.

More than 25 percent of the school days on our calendar are not full days of school.

So if it feels as though your children never have a normal week, you may be right.

The Prompt

The prompt below assumes you have Claude Cowork connected to your calendar, but that is not required.

You can also:

• Share a link to your school’s published calendar

• Upload a PDF or screenshot of the calendar

• Paste the important dates directly into the conversation

If your children attend different schools or school systems, include every calendar. Mismatched days are often where the biggest childcare problems, and some of the best travel opportunities, appear.

I wrote this as a prompt because that is where most of my readers are right now: in a chat window, providing background, and asking AI to help with one problem at a time.

That works.

If you have set up a Project, you are a step further along, since the context sits there and you are not starting cold with each request. But a Project holds background, not structure.

The next step is a real markdown setup, where the school calendar, the childcare gaps, the standing constraints, and your travel patterns live in files Claude reads every time. It’s a game-changer and once you see the value, you’re hooked. But you do not need to build the full system today. Start with the calendar.

Review the school calendar and my personal calendar for the 2026 – 2027 school year and build a condensed family planning reference I can print and use throughout the year. If my children attend different schools, treat each school’s calendar separately. Label every event with which school it belongs to, never assume a closure at one school applies to the other. Create a separate category for mismatch days, when one school is closed or dismissing early while another is in session. These dates may create childcare problems or travel opportunities.

FAMILY CONTEXT

• Children: [names, grades entering in fall]

• School(s): [school name]

• Childcare setup: [e.g., “Not enrolled in Extended Day — all early dismissal days require separate coverage” or “Enrolled in Extended Day Mon/Wed/Fri only”]

• Only include events that affect [grade range] students. Ignore events for other grades (graduation, placement tests, performances, etc.)

UPCOMING FAMILY COMMITMENTS

[List each with a specific date and what you need to know]

• [Brother’s/friend’s] wedding: [date range]. We want to arrive by [date]. Tell me how this overlaps with the school calendar and whether any adjacent closures let us extend the trip without missing school.

• [Family visit/trip]: [target month range]. Identify the best 1–2 week windows based on school breaks, PD days, and long weekends, and rank them by school days missed.

• [Planned trip, e.g., Florida]: [month]. What is the earliest realistic departure after school ends? Flag any conflicts with camps, activities, or family events.

Other important family goal: [event, deadline, or desired trip].

WHAT TO IDENTIFY

1. Every early dismissal, half day, PD day, and full school closure - with coverage implications given our childcare setup

2. Any mismatch days between the kids’ schools.

3. The 3-4 best travel windows of the year, including how many school days would be missed

4. The highest-risk childcare dates, ranked

5. Important school events that we may not want the children to miss, including field days, performances, class trips, or other special events listed on the calendar.

6. A short list of decisions that need to be made now vs. this fall vs. early next year

OUTPUT FORMAT

Deliver three things:

1. A condensed reference table: date, event, coverage needed or travel note. One row per date. No prose.

2. A short section for each upcoming family commitment: recommended dates, school days missed, tradeoffs, and the next decision I need to make.

3. A proposed list of 5–6 calendar reminder tasks with trigger dates - for my review before adding to calendar.

Keep the output tight enough to fit on 2 printed pages. Do not repeat calendar events as prose. Interpret the calendar and prioritize what matters.

TECHNICAL NOTE

For every event, use the date from the calendar entry itself. Do not calculate days of the week independently.

I still do not like how early school starts. I remain bewildered by the half days. And I would happily vote to restore the Memorial Day to Labor Day summer of my childhood.

But I cannot make the school year less complicated.

I can make sure the complicated parts do not sneak up on us.

I can solve the childcare gaps before they become emergencies. I can see where a long weekend might become a real trip. I can protect the wedding, the family visits, and the school events my kids would be crushed to miss.

The calendar does not get to decide what matters most to our family just because I failed to look ahead.

That is the little win: not organizing the school year perfectly, but making room for the life we want to have inside it.

Subscribe now

At the beginning of the summer, I had ambitious plans about the summer slide. Reading was an easy sell; math was not. It is hard to convince a child who does not love math that summer is the ideal time to do more of it.

So I started wondering if I could smuggle it in. Percentages, profits, valuations - maybe those land differently when they come wrapped in a pitch to Mr. Wonderful.

I opened a chat in Claude and typed:

My kids love Shark Tank. I would love to build a math game that they can play this summer that uses Shark Tank content to build skills in math. Thoughts?

I was asking for a plan.

I assumed we would map out the structure, decide what the game needed to do, and maybe write something official-sounding that I could take into Claude Code later.

Claude gave me three options: a quiz, a game where you play the investor, or a build-your-own-pitch experience. It recommended the investor version and asked one question: Did I want it in a browser?

I said yes.

Then it said:

Good. Let me build this now.

I genuinely did not understand what was happening.

I actually stopped and asked:

Are you actually building it? Or just specing it?

Building it.

A few minutes later, I had a working game and instructions to go test it.

No coding environment.

No setup.

No product requirements document.

I had a loose idea, picked one of three directions, and that was it.

Then my ten-year-old took over

The first version showed a kid-friendly pitch, gave you the numbers, and made you decide: invest or pass.

I played a few rounds and immediately wanted things changed.

The math was too hard for her. Some of the incentives made no sense - you could earn points for correct math without the math actually helping you make a better call, which is exactly backward.

We fixed it.

Then I emailed it to my daughter. She opened it, played for about a minute, and said, “This is so cool—I can’t believe you made this.”

Neither could I, frankly.

She also, within four minutes, had notes.

So we AirPlayed my laptop onto the TV and pulled up the game on one side and Claude on the other. She played. She told me what she wanted changed. She watched it change.

That was the part I did not plan for. She was doing math, sure. But she was also testing a product, giving feedback, and figuring out that the instructions she gave changed what came back. That is a thing I would have loved to learn at ten.

A few weeks later, we were at a friend’s house. She pulled it up on my laptop, and suddenly there were four kids crowded around a screen arguing about a deal.

What I had wrong

I had been sorting everything into two bins: things I could do in a chat and things that required me to go be technical somewhere else.

There is an enormous middle I was not taking advantage of.

You can build real, working, interactive things through a normal conversation - no coding required.

I want to be honest because plenty of times, the juice is just not worth the squeeze. I learned this through building trying to build my AI Super System.

This was not the same experience.

This was one conversation, immediate gratification, and a game my kid actually plays.

Once I saw it, I started seeing the category everywhere:

• A magazine quiz. An actual one. “Which Off Campus Character are You?” with a scoring key at the bottom, exactly like the ones we took in the back of YM while someone read the questions out loud by the pool.

• A states quiz for the kid who has learned exactly four states.

• A choose-your-own-adventure starring whatever your child is currently deranged about.

• A vacation countdown with trivia and a packing list, so someone other than you is the one asking whether the bathing suits are in the bag.

• A family game-show board for a birthday.

• A chore tracker your kids will love for nine days.

• A rough version of the idea you keep describing at dinner, before you spend real money finding out.

The trick is that it should be self-contained. No accounts. No database. No connecting to five other systems. Just one small thing that can reasonably live in one file.

One caveat: the easiest version of this lives on a laptop or desktop, where you download the file and double-click to open it in a browser.

Opening an interactive HTML file directly on an iPhone or iPad is not nearly as seamless. To make it reliably accessible from a phone or tablet - or to send someone a normal link - you would need to host it online using a service like Netlify.

I have not bothered, because the laptop version did exactly what we needed.

But you do not have to know any of that before you start.

You can just ask:

Could you build this as a single HTML file that I can download and open on my laptop?

That is the prompt.

I had made “building” so much bigger in my head than it needed to be. I had been budgeting weeks for something that took one conversation.

Less waiting to feel ready.

A daughter who thinks she is now both a product manager and VC. One weirdly good afternoon.

That is your little win.

We were talking about an expense for one of her kids - an unexpected trip abroad. It had landed right behind a charity event neither of us had budgeted for - the kind where you think you’re attending dinner and somehow end up bidding on several silent auction baskets you do not need.

At some point, one of us said it out loud.

I am hemorrhaging money.

I have written before about watching a tax bill leave my account and crying about it, even though I knew it was coming.

This was the same spiral, just a different invoice.

I understand that the discomfort of money leaving is not the same as money being a problem.

I know that - intellectually. But knowing it has never once stopped the emotional spiral.

About ten years ago, a favorite coworker handed me Morgan Housel’s The Psychology of Money. I loved it. If you know me, I have probably told you to read it.

I own most of Housel’s books and also love his blog posts. Highly recommend - he is so skilled at making complicated things feel obvious.

He is my go-to, not because I am looking for a financial advisor, but because when I spin out about money, what I actually want is someone to sit me down and remind me what game I’m playing and what it means to win.

I reference his thinking so often that I finally did something about it. I set up a project that exists for exactly one job: stopping the spin.

Or at least slowing it down long enough for me to find my brain again.

I didn’t upload the books. I told it whose thinking I trust, gave it the principles I actually return to, and told it what I needed it to be. Not a financial advisor. Not a cheerleader. The voice in the room that takes me back to first principles when I’ve lost the plot.

Now, when the spiral starts, I don’t doomscroll my own bank account. I open the project and just say it: Here’s everything hitting at once. I feel like I’m hemorrhaging money.

And it hands me back the plot. Not by reassuring me, but by asking better questions.

Is this money you set aside on purpose?

Are you confusing the discomfort of spending with something actually being wrong?

What did you say money was for, back when you weren’t in a panic?

That last one matters most.

Because I do actually have a philosophy about money.

I just forget it the second three camp invoices, two trips, and one worthy cause walk into my checking account at the same time.

For me, money is a tool for time, autonomy, and a little room to breathe.

The ability to be available to my children while they still live in my house.

The ability to take the trip while everyone still wants to come.

The ability to be generous without turning every yes into a personal crisis.

So if the money is leaving for the life I said I wanted, then the issue may not be the money. It may just be the discomfort of watching it move. That’s the whole thing.

I described what I needed once.

Now it is there whenever the chest-tightening starts.

And while money is the thing that does it for me, the structure is not really about money.

It is about catching yourself before the thought loop becomes a full-blown tornado.

The little win is not “use AI to manage your finances.” Please do not let AI manage your finances.

The little win is using AI to remember your own principles when your brain has temporarily misplaced them.

Go try it

You don’t need to own the books. You need to know whose thinking steadies you. A writer. A mentor. Your most level-headed friend. The person who reminds you that you are not actually on fire.

Open Claude or ChatGPT.

Start a project.

Set the instructions, something like this:

You are my money-panic thinking partner.

When I’m anxious or guilty about spending, don’t give financial advice, budgeting tactics, reassurance, or generic positivity. Help me recover my own money philosophy.

Draw on themes from Morgan Housel’s writing: money as psychology, independence, enough, risk, and the emotional side of financial decisions.

Ground me in these principles:

Money is a tool for autonomy, options, and time.

Wealth is the money I did not spend.

Money leaving on schedule is not a problem.

More only matters when I know what it is for.

Separate the financial fact from the emotional story.

Be calm, direct, and grounded. Name the loop, ask a few questions, remind me of the principle, and hand me back the plot.

Then, the next time your chest tightens over a number, go there before you go anywhere else.

One honest note. If you are going to borrow someone’s thinking through Claude or ChatGPT, pay the toll. Housel didn’t write those principles so a chatbot could hand them out for free. Buy the books. Let AI remind you. Let the human get paid.

And if you’re wondering whether this only works for money, it doesn’t. Money is just the thing that spins me out.

For you, it might be work. Or parenting. Or your inbox. Or your health. Or the thing you keep overthinking at 11:47 p.m. when you are too tired to be reasonable and too awake to let it go.

That’s the real little win here.

Not that AI solved my money anxiety.

It didn’t.

The win is that I built a place to go before the spiral picks up speed. A place that knows what I believe when I’m calm. A place that can hand it back to me when I forget.

So pick your spin.

Name the loop.

Build the project.

Then go there first.

That’s your little win.

By noon, I was on the couch. By 1:00, I was at the doctor. By 2:00, I had a pneumonia diagnosis. By 4:00, I was in bed - and I did not get up until the following Saturday.

I also never opened my laptop. Not once.

What I did instead: talked into my phone.

I’ve been using Wispr Flow for a few months now. It’s a voice dictation tool that works across apps - instead of typing, you speak. That probably sounds like a minor convenience. When you cannot breathe properly and sitting upright feels like a personal defeat, it is not minor. It is the difference between doing something and staring at the ceiling.

So from bed, I talked. To ChatGPT. To Claude. Back and forth, the way I always do - just without the sitting up part.

I asked Claude to read me my weekly brief out loud so I didn’t have to open my eyes. It did. Then I remembered something I’d been putting off for months: I had a spreadsheet in Google Drive with every camp the girls are attending this summer - along with dates, drop-off/pick-up times, locations, and friends - that I’d built during registration season and never actually moved into my calendar. I use ’s calendaring system, which means that all kid activities live in a dedicated “Kids” sub-calendar under my personal Gmail calendar. Getting it all in there had been sitting on my list since April.

I asked Claude to pull the file from Drive and create the calendar events. It did. Automatically, because I’d already set up CoWork to enable those kinds of actions. The camps are on the calendar now. I did not type a single event. It populated everything automatically.

Then I turned to ChatGPT.

This is where I need to explain something, because the two tools were doing very different things last week.

Claude is my operational brain. It runs the brief. It connects to my files. It executes. ChatGPT is something else - closer to an aesthetic one. And for what I needed next, aesthetic was exactly right.

If you’ve read the stylist post, you know part of how that happened. The Instagram accounts I follow, the style brief I built once and have never rewritten, the photos I’ve sent over time as I’ve tried to restyle rooms or figure out an outfit. ChatGPT has been living in my life for almost a year. It has seen my house. It knows my color preferences, what “printable” means in our home, and my precise threshold for whimsy. None of that happened in one sitting. It was built in small moments, over time.

What that means in practice is that when I finally picked up my phone to deal with summer, I didn't have to set anything up. I just had to ask. And what summer requires in our house is documentation.

My children are the children of two lawyers. They need it in writing. They like terms. They like exceptions. They like to argue about whether “clean up after yourself” applies if the towel is technically not on the floor but draped over a stool in a way that suggests future action. So yes, we need things in writing. Not because a sign will magically change anyone’s behavior - I am not new here - but because it removes me from the position of human reminder system and gives everyone something to point to besides my face.

When I said “I need a printable camp schedule - something cute,” I didn’t have to explain what cute means to me. When I said “make a bathroom sign,” I gave it a few items from the morning and night routines and asked it to finish the list. It did. Kitchen rules: when meals are open, when snacks are allowed, when the kitchen is closed. I said the framework. It filled it in.

I did not tweak any of it.

I hit print. Wirelessly, from my phone, from bed.

The output required nothing from me not because I got lucky with a prompt. It required nothing from me because of everything I had already provided to the tools, in different ways, over time. A stranger could have helped with some of this. But I would have had to explain everything. ChatGPT already knew. Claude already had access. I just asked.

This is the version of AI I wish more people were being shown. Not the futuristic one. Not the corporate workflow one. Just: you invest a little context over time, and when real life knocks you flat, you have something to reach for that already understands you.

The context is the gift you give your future self.

Here is the smallest way to start.

Open a new ChatGPT project and call it something boring and useful, like “Family Summer” or “House Stuff.” Then add a few notes you would otherwise have to explain every single time:

My kids are [ages].
Our summer pain points are mornings, camp logistics, wet towels, snacks, and screen time.
I like printables that are clean, cheerful, and not too cutesy.
I want help making signs, checklists, schedules, reminders, and instructions.
When you write for my kids, keep it short, friendly, and clear.

That is enough. Truly.

You are not building a command center. You are leaving breadcrumbs for the version of you who may one day need to make a bathroom sign from bed.

It doesn’t have to be elaborate. It just has to be consistent. A photo here. A preference there. A spreadsheet you built in February that you forgot about until you needed it. It compounds quietly, and then one day you’re horizontal with pneumonia and nothing falls apart.

You don’t need a perfect system. You just need enough already in place that when you’re operating on fumes, something can still get done. Because sometimes the future doesn’t arrive as a robot butler.

Sometimes it arrives as a grandmillenial bathroom sign you made from bed.

That’s your little win.

AI has a funny way of disguising itself as productivity. Nobody accidentally spends six hours playing tennis and convinces themselves they’re getting organized. Nobody reads three novels in a weekend and calls it workflow optimization. But reorganizing your digital infrastructure? Building the perfect system? That feels productive. Responsible, even.

The problem is that sometimes the setup quietly becomes the whole activity.

For the past two weeks, I’ve been deep in an AI spiral. Not the scary kind; the fun kind. The kind where every new thing you learn opens three more doors. The kind where you find yourself researching file structures on a beautiful June afternoon, sitting on your porch, telling yourself this technically still counts as being outside.

A lot of it is working. I’m building things I’ve wanted to build for years. AI is helping me think more clearly, write more quickly, and make things I never would have attempted on my own.

But then this past weekend, I ran headfirst into the wall.

If you’ve been here a while, you know I’ve been playing with Cowork - Anthropic’s desktop tool that lets you connect your apps, automate workflows, and build a little command center for your life. I genuinely believe in it and I think the sooner you start playing with it, the better. I’ll come back to why.

But this past week, I aimed too high. I decided to build a full Cowork operating system - not just for my personal life, but across my work account too. Everything connected, everything automated, one unified system. I’d been reading about it, testing pieces, mapping it out for weeks. I thought I was ready.

And to be fair, I did build it.

The problem was that I couldn’t get it working in a way I trusted. I ran into account security issues I couldn’t comfortably ignore, built workarounds that felt increasingly fragile, and spent hours organizing a file structure that depended on connections I didn’t actually have. By Sunday night I had a system that sort of worked, a list of things I still needed to fix, and exactly zero posts written.

Here’s what I learned: the elaborate version - full file mapping, cross-account automation - is genuinely complicated for a non-technical person. Not because Cowork is bad, but because the setup I wanted depends on connecting multiple accounts and that’s not something Cowork supports out of the box.

That’s a real limitation worth naming. But it’s also my limitation, not necessarily yours.

If your life mostly runs through one inbox and one calendar - whether that’s work or personal - you’re actually starting from a simpler place than I am. You can build something surprisingly useful without any of the workarounds I was wrestling with.

But here’s what I also learned, from everything that did work: when you give Cowork clear context and one specific thing to produce, it delivers. That part is not complicated. That part is available to you right now, this week, without any of the infrastructure I was breaking myself on.

And I didn’t post last week because I was too buried in the parts that weren’t working to write about the parts that were.

I stepped back and asked a simpler question.

What can I build today that’s useful right now?

Not for some future version of my life where everything is perfectly connected and elegantly automated. For my actual life. This week. With real kids, real deadlines, and a very real number of hours in a day.

The answer wasn’t a complex automation or a custom workflow or an AI operating system. It was this: I want to stop walking into Monday with no idea what the week actually holds. The appointments, the pickups, the activities - the things that quietly ambush me at 4pm if I don’t see them coming.

So I built the one thing I actually needed and could build relatively quickly. It took a context file and one good prompt. No elaborate file mapping. No cross-account security drama. And it works.

I turned it into a toolkit you can use too - and more importantly, it’s proof of what Cowork can do for you right now, before you’re tempted to over-engineer it the way I did.

Here’s what it produces

Before the how, the what. This is a real brief I generated for myself last week - sanitized and condensed, but real:

📋 Weekly Family Brief — Week of June 9

① WEEK LOAD: PACKED ⚠️ A heavy, irregular week. A trip (Thursday–Sunday) sitting on top of the last week of school — two early 11:10am dismissals and the last day of school Tuesday. Trip + early dismissals + end-of-year events is exactly the kind of pileup that catches me off guard. Downtime is thin; protect what’s left.

② DO THIS WEEK 🔴 Pay Q2 estimated taxes (Federal + state) by Monday June 15. I’ll be traveling, so do this before I leave Thursday or it slips. 🔴 The girls are coming on the trip. This isn’t a childcare problem — it’s a packing one. Pull together what they’ll need for the days we’re away. 🟡 Decide on the graduation event Wednesday. Decide now, not day-of. 🟡 Book Pilates and strength for the next two weeks. Classes fill early — reserve now. 🟢 Father’s Day is two weeks out. Tuckernuck is running a sale now.

③ DETAIL

❤️ Marriage Date-night chain: nothing booked. With the trip, the realistic window is after we’re home Sunday or early the following week. None of the three pieces — time blocked, sitter, plan — exists yet. Hubs’ travel calendar: nothing logged this week.

🏃 Me Booked: Friday tennis (7:45am), golf Sunday, tennis again the following Friday. Gap: zero Pilates and zero strength sessions in the next two weeks. That’s my 2-and-2 target, and it’s the thing that silently drops on a travel week. Book before classes fill. Also: dentist next Wednesday 8am, and a reminder to call about a lab order.

That’s it. Two minutes to read. Everything I need, nothing I don’t.

Here’s the part I want you to notice. The first time I ran it, it got the trip wrong. It saw flights on my calendar and flagged a childcare scramble - who’s covering the early dismissal while you’re away? But the girls were coming with me. So I told it once: when a trip is also on the girls’ calendar, they’re traveling too. Now it never gets that wrong. It flags packing, not babysitters.

That’s the whole magic, and it’s not magic at all. It knows my kids are coming because I told it. It knows which emails are noise and which I care about because I told it. It knows my exercise targets and my tax deadlines because I told it - once. And now it remembers.

The Toolkit

Three parts. You do Part 1 once. You set up Part 2 once. Part 3 is where it becomes a permanent fixture instead of a thing you rebuild every week.

What you’ll need: a paid Claude plan, and your Google Calendar and Gmail connected to Claude. This is the one piece of setup that actually matters - the brief can only see what you’ve connected. Connect Google once and it works everywhere: in a regular chat and in Cowork. (If you use Outlook or Apple Calendar, Part 1 still works, but the connections in Part 3 are written for Google, so you’d adapt.)

If you haven’t connected anything yet, start with my Claude + Cowork quick setup guide. Make sure training mode is turned “off” and then come back.

PART 1 — The Interview Prompt

Do this once. 10–15 minutes.

This is the part most guides skip. Paste this into a new Claude chat and answer one question at a time - it only asks one thing at a time, so it never gets overwhelming. At the end it writes you a setup file. That file is the thing that makes every future brief personalized to you and what you care about.

You are my personal chief of staff, helping me design a Weekly Family Brief - a Sunday-evening look-ahead that turns my scattered calendar and inbox into one clear picture of the week, so I carry less of it in my head.

Your job right now is not to write the brief. It is to interview me and learn what my ideal brief should contain. I’m most likely someone juggling family life and work, so you can start there - but check your assumptions with me and adjust if my life doesn’t fit that frame.

How to interview me:

• Ask one question at a time. Wait for my answer before the next.

• Be warm and patient. Briefly explain why you’re asking when it helps.

• Don’t assume the categories that matter to me - discover them by asking.

• When I’m vague, gently ask me to get specific rather than guessing.

• Keep going until you could confidently design a brief tailored to me.

Cover, through natural conversation:

• When I’ll read the brief, and who reads it (just me, or a partner too).

• Where things fall through the cracks — my “oh no, I forgot” moments.

• The parts of life that matter most: family, kids, relationship, my own health and downtime, any side project or business.

• Work: ask whether I want work included or kept out, and whether it lives on the same calendar as my personal life or a separate one. Don’t assume.

• Which calendars and which email account you should review.

• What gets sacrificed when my week fills up, that I want protected.

• What creates stress when it’s overlooked.

• What I want you to ignore - junk email, certain senders, noise.

• The format I want: how detailed, how organized, how I like to read.

When the interview is complete, produce one document in clean Markdown titled “Family Brief — Operating Manual.” This is the durable memory for my brief. Fill in every section from what I told you, following this skeleton:

# Family Brief — Operating Manual

Owner, time zone, when I read it, who reads it.

## Scope

- Which calendar(s) and email to review; whether work is in or out.

## Calendar Map

- Each calendar I named and what it holds.

## Travel Logic

- How to tell who’s on a trip. (A trip that also appears on my kids’

calendar means the kids are with me — so flag packing, not childcare.)

When unsure, ask rather than assume.

## Standing Rules

- My recurring must-knows (early dismissals; exercise targets and how

far ahead to check classes; protect downtime; date-night logistics).

- VERBATIM: “Print each event’s weekday from its own date field. Never

calculate the day of the week yourself.”

## Content / Side Project

- What to check, what to leave alone.

## Email Filter

- What to surface (actions, deadlines, RSVPs); what to suppress; my

“always show me” whitelist of brands or senders.

## Output Format

1. Week Load verdict — Packed / Manageable / Light, plus one line.

2. “Do This Week” — one ranked action list (action + why it matters now).

3. Detail — short bullets by life area.

End by telling me to save this document as family-brief-context.md - I’ll need it for the next step. Ready? Ask me your first question.

When it’s done, save the file exactly as it tells you: family-brief-context.md. That filename matters - the next prompt looks for it by name.

PART 2 — The Brief Prompt

Test it once in a chat to see it work.

This is the prompt that generates your brief. To test it right now: open a new Claude chat, attach the family-brief-context.md file you just made, make sure your Google connections are on, and paste this:

Read my attached family-brief-context.md and follow every rule in it. Then generate my Weekly Family Brief for the next 7 days, with a 2-week lookahead for exercise and any big items that need planning, using the calendar(s) and email named in my file. Follow my file on whether work is included or kept out.

Apply the reading rules from my file: print each event’s weekday from its own date field (never calculate it); use my travel logic to figure out who’s on a trip, and don’t raise childcare flags when my kids are traveling with me; if something is ambiguous, note it and ask rather than guessing.

Structure the brief exactly as my file’s Output Format section says: Week Load (Packed / Manageable / Light, plus one line); Do This Week (one ranked action list — action + why it’s time-sensitive); Detail (short bullets by life area).

Keep it scannable and direct. Show me the full brief right here.

Read it like it’s Sunday night. See what it nails and what it misses. When something’s off, tell Claude and ask it to adjust the underlying file. The file is the part you tune. (That’s exactly how I taught mine about the trip.)

PART 3 — Give It a Permanent Home in Cowork

This is what makes it last.

Testing in a chat is great for seeing it work. But you don’t want to re-upload your file to a fresh chat every Sunday for the rest of your life. That’s where a Cowork project comes in - a dedicated workspace that holds your file, your instructions, and your connections in one place, so running your brief becomes a single step.

1. Open Claude Desktop and create a new Cowork project. Call it “Family Brief.” (Projects live in the desktop app, and they keep their own files and memory — which is exactly what you want for something you’ll run weekly.)

2. Add your family-brief-context.md to the project. Now it lives there permanently. No more attaching it each time.

3. Confirm Google Calendar and Gmail are connected in the desktop app’s settings. Because Google connects across all of Claude, if you did this for the Part 2 test, you’re already done.

4. Run the Part 2 prompt inside the project. Since the file already lives there, you can shorten it to: “Generate my Weekly Family Brief following my context file.” Read it. Tune the file if needed.

5. Schedule to run every week. Open a chat in the CoWork project and simply type, “Schedule task to run every Sunday at 7 pm.”

That’s it! Every Sunday, open the project and get your brief.

Your little win this week

Open a new Claude chat, paste the interview prompt, answer the questions one at a time. By the end you’ll have a setup file built around your kids, your calendars, your version of a week that works.

If you want to keep going: test it with Part 2, watch it work, then give it a home in Cowork with Part 3. But even if you stop after the interview, you’ll have done the part that matters.

One last thing - the family brief is just one version of this. The same approach builds a work brief, a travel-prep brief, a side-project brief - whatever your version of “I need to stop walking in blind” looks like. The interview prompt in Part 1 is the reusable engine. Point it at a different part of your life and it builds a brief for that instead. The structure stays the same. The output is whatever you actually need.

You don’t need a folder structure. You don’t need cross-account automation. You don’t need the elaborate version I spent a week breaking myself on.

You need twenty minutes and a paid Claude account.

Less scrambling Monday morning. A clearer Sunday night. More of the week actually going according to something resembling a plan.

That’s your little win.

I didn’t know that until last weekend.

A ChatGPT project I’d been running for about three days handed me the label, along with a description of what it meant:

Your ideal book contains:

• one emotionally constipated family

• at least one summer house with history in the walls

• complicated women quietly ruining each other’s lives out of love

• old New England money OR aspirational Boston Irish energy

• intergenerational guilt

• restrained yearning

• weathered hydrangeas

• somebody making chowder while suppressing a breakdown

It also handed me a list titled “Books You’ll Probably Never Fully Buy Into” which included:

• Tomorrow, and Tomorrow, and Tomorrow

• Remarkably Bright Creatures

• Malibu Rising

• most “hilarious millennial woman figuring it out in Brooklyn” novels

• dark prestige misery novels that mistake suffering for depth

This felt like every magazine quiz I ever took in middle school - flipping through YM and Seventeen to find out what kind of girlfriend you were, or which Spice Girl matched your study habits. You’d circle A, B, or C and turn to page 47 to confirm what you already suspected about yourself.

Those quizzes only ever told you what you already knew. This one told me something I didn’t.

Here’s how this started.

Last Wednesday a girlfriend from college came over for dinner. She left around eleven and it was pouring. The sun has not come out since. The weekend that was supposed to be the first real pool weekend of the summer became three grey days of gentle existential drift, and I spent most of it inside, reading - and, more honestly, thinking about reading.

My kids can finally be in the pool by themselves. I am in the era of sitting in the sun with a book while children I love splash safely twenty feet away. This is how I want to spend every evening this summer. I started building my list.

And then I hit the wall I hit every summer.

Everyone is recommending things. The group text. Instagram. The friend at dinner. The mom at pickup who said, “you have to read this”. And I have learned - the hard way, multiple times - that other people’s books are not always my books.

A few summers ago, every woman I knew was reading Daisy Jones & The Six. I could not get past page thirty. The dialogue format was killing me. I never finished it. (For what it’s worth, I loved the show.) Same with The Seven Husbands of Evelyn Hugo. Couldn’t do it. So maybe Taylor Jenkins Reid just isn’t for me - but that is exactly the point. I had to read try to read two books I didn’t enjoy to figure that out.

My friend mentioned she’d been running a content curator in ChatGPT for a while. The part that stuck with me wasn’t that she gets great recommendations - it was that she had learned more about her own reading taste than she knew before. She could now articulate what she wanted from a book in a way she never could before.

That’s the part that got me. Not better recommendations. Better self-knowledge.

So Saturday morning, between rain showers, I set one up for myself.

Meet the Book Club of One

This is the next member of your AI Dream Team - joining the AI Stylist and the Weekend Concierge. Same logic: set up a dedicated project space once, talk to it like an actual person who knows you, and then never start from scratch again.

Quick note on tooling. I built this one in ChatGPT, not Claude. I’ve been preaching Claude for execution work - when you want something done (build the calendar blocks, draft the emails, run an actual task), Claude is better. But this isn’t execution. This is curation and conversation, and ChatGPT is genuinely great at it. Its slightly eager, friendly demeanor - which can be irritating when you’re trying to do critical analysis - is exactly what you want in this instance.

How I Built It

There are three parts to this and none of them are hard.

1. The Project Instructions

I created a Project in ChatGPT called Reading Rainbow. (LeVar Burton, forever.) Then I went into the Project Settings and pasted in a set of Instructions - the brief that tells the project who it is and how to behave every time I open a new chat.

This is the part most people skip. Most people open ChatGPT, type recommend me a book for the beach, get a generic answer, and quit. Instructions are how you turn a generic chatbot into something that actually knows you. They load into every conversation automatically. You write them once.

Here’s what I used. Steal it.

You are a personalized reading curator and literary taste profiler. Continuously refine my reading profile based on books I loved, liked, disliked, abandoned, or felt mixed about.

Track: genre and subgenre preferences, writing style and prose, pacing and complexity, emotional tone and themes, character preferences, literary vs. commercial balance, humor, darkness, romance, sentimentality, and common dealbreakers.

Support both fiction and nonfiction. For fiction, pay attention to prose, atmosphere, pacing, character depth, themes, and emotional impact. For nonfiction, pay attention to subject matter, insight level, argument style, narrative quality, practicality, and complexity.

When recommending books: prioritize taste alignment over popularity. Recommend bestsellers when they genuinely fit. Mix newer releases, overlooked gems, and strong backlist titles. Explain recommendations casually and specifically. Include occasional “stretch picks.” Respectfully explain books I may not enjoy.

When asked for a reading list, provide: a short Reader Personality profile, recommended books with title, author, why it fits, vibe tags, and best-for context, and popular books that may not resonate with short taste-based explanations.

Tone: casual, smart, perceptive, and fun - like a well-read friend with excellent memory for someone’s taste.

That’s the entire brief. Paste it into the Project Instructions within the Project Settings.

2. The Bookshelf Voice Notes

This is the process I want you to copy exactly.

I did not sit down at my laptop. I did not type up a thoughtful list. I did not even interview.

I opened ChatGPT on my phone, hit the microphone button, walked over to my bookshelf, and started talking. This is the actual transcript from Chat:

Loved Saints for All Occasions by Courtney Sullivan — loved the family drama, the New England setting, the trauma and tragedy, the Catholic hints. Loved Maine — the evolution of a home as the center of a family, the WWII background, the big Irish family drama. Circe — liked — loved the Greek mythology, but it dragged at times. Loved Elin Hilderbrand’s Five Star Weekend and 28 Summers — perfect beach reads, female friendships, Nantucket setting. Disliked Daisy Jones and the Six — hated the dialogue. Mixed on The Cliffs — thought I’d love it because I love real estate and old houses, but the supernatural element turned me off. Abandoned Demon Copperhead — too depressing, unlikable characters. Cutting for Stone — top favorite — love the medical element, the relationships, the setting in Africa. Lessons in Chemistry was OK. Felt a little canned.

Shorthand. Off the cuff. Fiction first, because that’s what was on the shelf.

The next day I did the same thing with my Audible library - mostly nonfiction.

Loved When Breath Becomes Air. Loved The Courage to Be Disliked. Tried The Anxious Generation but couldn’t finish — too depressing on audio, maybe better in print. The Friday Afternoon Club by Griffin Dunne — good listen, liked the family drama intersected with real LA history. Greenlights — fun biography. Loved the Walter Isaacson book on Elon Musk — super interesting human, listened end to end.

If you’ve never used ChatGPT in voice mode, this is the use case that will sell you on it. Typing this kind of brain dump is brutal. Talking it is effortless. And the project listens, asks the occasional follow-up, and starts assembling a picture of you.

3. Keep Feeding It

Here’s the part that turned this from a fun afternoon into an actually useful tool.

Over the next few days, every time I thought of another book I’d read - a college favorite, a book I’d quit and forgotten, an audiobook I’d half-listened to on a road trip - I opened the project and added it. New chat, new entry, more data.

Then I ran the reader profile again.

The first version made me laugh outloud. The second version was more specific. The fifth version actually gave me direction. It pointed me toward authors I’d never heard of. It told me which books on my list to skip. It explained why I’d loved one book everyone hated and bounced off three books everyone loved.

It remembers everything. It pushes back. When I mentioned a book I was thinking about reading because everyone was talking about it, it told me I’d probably hate it and explained why, citing patterns from books I’d quit. I checked the reviews of the readers I actually trust. It was right.

I would have read forty pages and given up. Instead, I’m reading something that’s been on my list for two years and I’m completely in it.

What I Didn’t Expect

I thought I was building a recommendation tool. What I built was a self-knowledge tool.

I now know that I read for place. That I will tolerate slow pacing if the atmosphere earns it. That I bounce out of any book where the dialogue feels stylized or “writerly.” That I cannot do trauma-as-spectacle. That a house can be the most important character in a book for me and that this is, in fact, fine.

I have spent a lifetime reading and never named any of this out loud.

Knowing it changes how I shop for books. How I take recommendations. How quickly I decide to quit something that isn’t working. How confidently I say that’s not for me without feeling like I’m missing out on something everyone else loves.

Try This Tonight

Open ChatGPT. Create a new project. Call it whatever you want - Book Club of One, The Reading Room, your name’s Library.

Open the chat, tap the microphone, walk to your bookshelf, and start talking.

“I’m going to tell you about books I’ve read. Loved, hated, quit, listened to, would re-read, would never finish if I had to read it again. I’m doing this in shorthand. Just listen and remember. We’ll build my reader profile at the end.”

Then just talk. Fiction, nonfiction, what was on your nightstand last summer, what’s been on your TBR list for three years that you keep meaning to read.

Then keep feeding it for a few days. The first profile is the joke. The fifth version is the tool.

Fewer books you quit at page twenty. More evenings actually spent reading instead of scrolling for what to read next. A clearer sense of what’s actually yours.

That’s your little win.

Drop your reader profile name in the comments. I'll go first: Coastal Catholic Family Saga Maximalist, at your service.

A few months ago I reached out to a friend and client because I felt like an AI failure.

I had this app idea. Specific, fully formed, something I genuinely believed people would use. I’d been turning it over in my head for two years. I’d experimented on my own and knew enough to know I wasn’t getting it off the ground. I was reaching out for the name of someone I could pay to build it for me.

Instead, he offered to teach me himself. In person.

He was in Miami. It was cold and snowy in DC. I booked the flight immediately.

A week later, I sat cross-legged on a living room floor eating sushi and building with someone who was genuinely excited about the thing he was teaching me. I’d been using ChatGPT for a while. I’d experimented, unsuccessfully with Replit. I tried to ask ChatGPT to walk me through the build - also unsuccessfully. That day, for the first time, I opened Terminal on my computer. And by the end of the afternoon I had the beginnings of the app I’d been telling myself I couldn’t build.

I came home and could not stop thinking about the distance between what I’d just done and what I’d been telling myself I couldn’t do.

I am building the thing that lived in my head for years while I waited to feel ready. I have no idea if it will work. But I am actually building it.

I am a builder. I love saying that.

That’s what moving from the other side of the table to this side feels like. I want that for you too, whatever your version of it is.

A month later, on a Wednesday morning, I started this newsletter.

Not to force myself to keep going. To stop privately overthinking it and start doing it out loud.

I’m an attorney. Overthinking is the job. Hitting publish was the part I had to teach myself.

That was ten posts ago. And what happened in Miami is something I’ve now started noticing everywhere: the people who become comfortable with AI usually don’t start in high-pressure environments. They start in the low-stakes corners of their actual lives.

I see it in my clients, in my friends, in the women in my life who keep telling me they should probably learn this stuff “eventually.”

I spend my professional life advising companies on AI. I spend the rest of it watching capable women quietly conclude they’re probably behind already.

Here’s the part nobody really says out loud: most employees have no real incentive to get good at the tool they suspect is being measured against them. That’s not resistance. It’s rational. People don’t eagerly master the thing they’re worried might replace them. So they experiment carefully - a summary here, a draft there - inside the most monitored, highest-stakes environment they will ever encounter the technology in.

Then leadership concludes the problem is adoption.

I don’t think the problem is adoption. I think it’s the setting.

The people I’ve watched actually become fluent didn’t get there at work. They got there in their own lives. Planning trips. Organizing schedules. Untangling logistical chaos. Solving ordinary annoying problems where the stakes were low and nobody was evaluating them. They learned because they were relaxed enough to be curious. And then they brought that confidence back into work with them.

This gap is widest with women. Not because of capability - I’d argue the opposite. But the way the AI industry talks about itself signals pretty clearly who it thinks the audience is. A lot of very capable women have correctly read that signal and quietly opted out.

Meanwhile, those same women are already doing the kind of work AI happens to be remarkably good at supporting. Coordination. Logistics. Research. Planning. Context switching. Remembering everything for everyone.

The mom job may be the one job AI isn’t taking.

But it turns out to be a very good place to learn how to use it.

That’s the bet I’m making with this newsletter.

Not that you should become an AI expert. Not that you should optimize yourself into a productivity robot. Just that if you build enough comfort with these tools in the low-stakes corners of your actual life, something shifts. The intimidation wears off. The confidence follows you back into everything else.

Less managing. More living.

That’s the little win.

If you’re new here - welcome. I write one of these every week. If you’ve been telling yourself you’ll get to AI “eventually” - start with the Blueprint pinned to the top of the publication. Fifteen minutes, no commitment. Otherwise the only thing I ask is that you try something this week. The smallest thing on your list. Tell me how it goes.

The Claude CoWork Blueprint is a quick guide. Fifteen minutes, start to finish. It walks you through getting set up the right way - privacy settings handled, Gmail connected with permissions that actually make sense, no overwhelm.

It’s the version of this I wish someone had handed me when I started.

A Little Win · Claude Blueprint

895KB ∙ PDF file

Download

Download

Download the full two page Blueprint above. Then go try one thing.

So when the first Saturday morning after the opening ceremonies rolled around and the girls were snuggled into bed, we turned on the coverage to keep the party going and got… cross country skiing. Bleh - not exactly the heart-pounding Olympics excitement we’d been promised.

At an ungodly hour, fueled entirely by Olympic delusion, I bought Peacock because the marketing promised I would never miss another event again. Any event. All events. On demand. We were not going to miss a single QuadGod performance.

The app was terrible.

I watched three episodes of the Scrubs reboot post-Olympics, then forgot Peacock existed entirely, and confidently assumed I had canceled it.

I had not.

It had been quietly billing me for the better part of six months.

I found this out because I handed my inbox to Claude.

Not metaphorically. Literally.

I was not looking for it. I was trying to build a label system.

Here’s where we are in this series. The first issue was the purge - 394,196 unread emails, the Swedish Death Clean approach, archive the backlog, cut off the flood at the source. That part I did alone.

This week I handed the clean inbox to Claude. Specifically to Claude CoWork, which is Claude with a direct connection to your Gmail - so the conversation doesn’t just give you advice, it takes action. Creates labels. Runs searches. Reads your inbox and surfaces what it finds.

If you're not set up yet, I put together a free guide - the Claude CoWork Blueprint - that walks you through the whole thing in about fifteen minutes. It's inside this post - subscribe free to get it. Do that first, then come back here.

Subscribe now

I came in with a plan. Ten labels, already decided. I am a person who has opinions about everything - including organization. I had a vision.

Claude had questions.

Not in an annoying way. More like a good advisor who knows that the system you walked in with is the system you built without all the necessary information. It wanted specifics. Not just “Medical” - which doctors, which labs, which pharmacy, which patient portal. The more precise the inputs now, it explained, the more the labels would actually work when we set up filters later.

And then it said something that completely changed how I thought about organizing email:

Keep the labels broad.

You do not need twelve nested folders for tax documents anymore. You need one Financial label and the ability to search intelligently later.

The more complicated our systems used to be, the more we were compensating for doing everything manually.

That reframe alone was worth the session.

Then it asked if it could scan the inbox - not something I’d requested, just something it offered. Said it could search for recurring billing patterns and surface senders I hadn’t thought to name. I said yes.

What came back was a precise picture of everything billing me regularly. Most of it expected. And then, near the bottom of the list, Peacock TV. Via Amazon.

I was annoyed. I was a little embarrassed. And then I canceled it, which I should have done seven months ago when I realized cross country skiing was not the content I was promised. I enjoyed the brief comedic reunion with JD and Turk, but not enough to pay $10.99 / month.

But here’s what surprised me most: you don’t have to wait for filters to get value out of this.

Filters and automations - that’s Post 3. But before a single filter was in place, I started asking CoWork to just read my inbox and tell me what it found. Real questions about real life.

So I asked.

Here’s the exact prompt I used:

And within minutes, I had a complete operational briefing for my household:

Action items (May 13–27, 2026)

DO / SHOW UP

Thu May 15

• Charlotte/Grace pick up Liam and Owen Foster from aftercare today (Megan Foster confirmed last week - you said yes).

• Choir rehearsal 3:10–4:10pm.

• Track practice 5–6pm.

• Spotlight Dance Tumble Showcase 6:55pm - parents invited for last 20 min of class.

• PTO meeting 7pm in the Sunnyvale library (optional - agenda linked in the email).

Fri May 15

• Spotlight Cheer dress rehearsal 6:15–6:30pm at the local high school.

Sat May 16

• Track meet (Grace). Elementary arrival ~1:20pm. Bring: water, sunscreen, snacks (PB&Js, fruit, granola bars). Mid-80s forecast.

• Spotlight Cheer recital pictures 9am + recital showcase 10:30am, local high school. Likely a scheduling conflict with the track meet - flag for Hubs.

Sun May 17 (tentative)

• Lacrosse playoff, time TBD at the away field. Coach Lauren will confirm in the app. Grace playing goalie - she’ll have both team helmets; you said you’d also try your brother’s old one at home.

Thu May 21

• No choir rehearsal.

Fri May 22

• Field Day; Early dismissal.

BUY / PREPARE

• State brochure project (Charlotte, 3rd grade) - last project of the year. Mrs. Mitchell sent directions home; the brochure paper comes home Monday 5/11. Confirm Charlotte has it and start planning over the next two weekends.

• Division facts practice (Charlotte) - Mrs. Mitchell flagged this is tripping students up in fractions.

• Track meet kit (Grace, Sat 5/16) - water bottle, sunscreen, hat/shade, healthy graze snacks, sit-on-the-ground gear.

• Lacrosse goalie helmet (Grace, Sun 5/17) - try your brother’s old helmet at home this week as a backup; Coach Lauren will bring both team helmets.

• Cheer recital costume + hair/makeup prep (Grace, Sat 5/16) - confirm what Spotlight wants.

Money - Pay / Resolve

• Bridgemath Academy - Next charge June 1 for 2026–2027 (both girls). Resolve.

• Crescendo Music - Autopay failed for installment 9/10 (both girls, Symphony at Sunnyvale). Update card and re-run.

• End-of-year class gift for Mrs. Mitchell + Mrs. Bennett - due Fri May 29. Venmo @[room parent] or PayPal, note “Collection for Mitchell.”

I also asked whether I needed to buy anything for Field Day.

It checked, found no outfit email yet, told me what last year’s pattern suggested, and said to wait before buying anything.

I asked about active sales from stores I actually shop. It pulled what was live, flagged what had expired, and told me the Tuckernuck sale was ending soon.

I asked about returns from the last two weeks - specifically whether the refunds had actually hit my account. It found two, told me the amounts, and flagged which one to chase if it hadn’t posted yet.

No filters. No automation. Just a clean inbox, a connected tool, and the ability to ask a real question about my actual life and get a useful answer the same afternoon.

The labels matter too - not just for today, but because without structure, you’re six months away from 394,000 unread emails again. The labels are what keep the volume from coming back. And they’re the foundation the filters will run on when we get there.

But you don’t have to wait for the filters to feel the difference.

Your prompt to start.

If you want to run the label interview yourself - no Gmail connection required, works in regular Claude - start here:

The interview is the work. What it surfaces is almost always more complete than what you walked in with.

Twelve labels are live. Peacock is canceled.

Next week: the filters go in, and things start running without you.

That’s your little win - and it cost me exactly one conversation.

P.S. If you’re ready to set up Claude CoWork, I put together a free setup guide - the Claude CoWork Blueprint. Fifteen minutes, do it on your laptop.

A Little Win · Claude Cowork Blueprint

925KB ∙ PDF file

Download

Quick 15 minute guide to setting up Claude CoWork.

Download

Share

I was at a school gala on Saturday night when I heard it. Again.

My wife is amazing. She does everything. I don’t know how she manages it all.

And I smiled. Because he meant it. The admiration was completely real. These are good men who love their wives and genuinely can’t believe what they’re watching.

But here’s what I always think when I hear it.

She knows. She’s there. She’s the one actually doing it.

And the speech - as sweet as it is - isn’t actually helping.

Here’s the thing about the woman in your life. She’s a superwoman. Full stop. And AI - when she actually has access to the good version of it - is like handing a superwoman a superpower.

But we’re getting ahead of ourselves. Let me paint you a picture of her Tuesday first.

She’s up before anyone else. She’s already thinking about the thing she forgot to follow up on, the appointment that needs to be rescheduled, the work deadline sitting at the edge of her morning. She gets everyone where they need to be. She shows up to her job - the one she cares about, the one she’s actually good at - and she delivers. She manages the group text. She checks in on her mother in law. And then somewhere in the middle of a perfectly orchestrated Tuesday, the kid is sick. Or the plumber didn’t show. Or she’s on the side of the road with a flat tire, on a call.

Handled.

And then - because she is who she is - she fits in the workout. The strength training she’s committed to. She’s hitting her protein. She’s working on her tennis and golf games, genuinely, with real commitment. She’s trying to drink more water. She is doing the things that are just for her, carved out of a day that doesn’t really have room for them.

She is running a very complicated operation. And she makes it look like nothing.

Here’s what she would never say out loud: the speech doesn’t help. The I don’t know how you do it - she hears it, she loves you for saying it, and then she goes to bed that night still holding everything she was holding before. The compliment lands and then the to-do list is still there.

What she actually needs isn’t more words. It’s bandwidth.

That’s where you come in.

Open your phone. Go to Claude.ai. Hit upgrade. Get her the Max plan, or at the very least, the Pro plan.

Yes, really. That’s the gift.

It’s not going to do her life for her. She doesn’t need that, and honestly she wouldn’t want it. But it’s going to corral the school emails and give her an action plan. It’s going to cut through the noise in the moms’ group text and tell her what actually matters. It’s going to write the emails - the ones she’s rewritten four times because she wants them to be right - in a fraction of the time. It’s going to help her prioritize the calendar. Not her calendar. The calendar. The one she’s running for every single person in this family, in her head, at all times.

So she can finally put that energy toward the stuff that’s actually hers.

She hasn’t set this up for herself because she’s too busy setting everything else up for everyone else. She’ll get to it. She keeps meaning to. It’s on the list, somewhere between the dentist appointment and the permission slip.

I’ll be honest - I don’t know exactly what she’s going to do with it. I just know what it did for me. And I know this: if someone told you they had the best nanny, the best cleaner, the best babysitter in town - you wouldn’t sit back and think about it. You’d get that number before they left the driveway.

This is that call. Don’t let her miss it.

You don’t need to wrap it. You don’t need a card. You don’t need to explain it. Just set it up, hand her the login, and tell her it’s hers.

Do not even think of buying that cheap scented candle from TJ Maxx.

Get her the superpower. She’s already the superwoman.

My daughter missed a birthday party last year because I never saw the invitation.

It was in my inbox. It had been there for two weeks. Buried somewhere under promotional emails, app notifications, and things I’d been meaning to get to since 2020. By the time I found it - running the search I now run every single week - the party was over, my kid was devastated, and I was composing the most awkward “I’m so sorry we missed it” text to a mom I wasn’t even sure I’d ever met in person.

Here’s how bad it had gotten: I have a recurring block on my calendar every week where I search my inbox for evites. From Evite, Paperless Post, Punchbowl, anything with “birthday” in the subject line. Not because I’m organized. Because I’ve learned the hard way that if I don’t go looking, I’ll miss it. That is the workaround I built instead of fixing the actual problem.

The actual problem is that I have 394,196 unread emails in my personal inbox. And that’s not counting my work inboxes.

I’d been ignoring that number for years. Watching it climb, telling myself I’d deal with it, and then not dealing with it. It became background noise - until it became a missed birthday party. And standing there trying to explain to my daughter why we weren’t going, I thought: okay. Something has to change. Not the workaround. The actual thing.

And then I thought about Swedish death cleaning.

If you haven’t encountered it yet: it’s the Scandinavian practice of ruthlessly clearing out your possessions - not to spark joy, but to get rid of what doesn’t belong before it becomes someone else’s problem. No sentimentality. No sorting. Just: does this have a place in my life or doesn’t it? If not, gone.

For an inbox with 394K+ emails, that felt exactly right.

That’s what I did with my inbox. And that’s what we’re doing here.

I didn’t bring AI into this quite yet. This is the part you just have to see clearly first. But I knew where this was going - because once this was clean, I wasn’t planning to manage my inbox anymore.

Full transparency: this did not take me twenty minutes. I didn’t do this in one heroic Saturday either. I’ve been chipping away at it in small pockets - coffee, school pickup wait, the ten minutes before I open my laptop. If your inbox is as far gone as mine -that’s the move. Not a heroic Saturday afternoon. A daily session, like a workout, until the weight is gone.

This is Part 1 of four. By the end of this series, most of what we’re doing manually right now will run without you. But you don’t start with a system. You start with a purge. So that’s where we’re starting too.

Before you touch anything — Know what you’re keeping

Before you open a single unsubscribe view, be honest with yourself about two things: what you can’t afford to lose, and what you actually want. Everything else goes.

Financial Banks and credit cards · Investment and brokerage accounts · Insurance (health, home, auto, life) · Mortgage and utilities · PayPal, Venmo, Zelle · Tax and accounting · Anything you’re actively billed by

Health Doctor, dentist, specialists · Pharmacy · Lab results · Health insurance

Family and school Your spouse or partner · Immediate family · Your kids’ school, teachers, coaches · Pediatrician · Sports leagues and team communications · Camp and after-school programs

Legal and government Lawyers · IRS and state tax · HOA · Passport and visa correspondence · Jury duty

Travel you actually use Frequent flyer programs · Hotel loyalty · TSA PreCheck / Global Entry

The ones you actually want The retailer you genuinely shop at. The newsletter you open every time. The brand that feels like a treat, not noise. You know which ones these are. Keep them. You’re not unsubscribing from Sephora. You’re not a monster.

Everything that doesn’t make either list? That’s what we’re clearing.

Step 1 — Cut off the flood at the source

Before you touch the backlog, stop what’s coming in. Otherwise you’re bailing a sinking boat without plugging the hole first.

Both Gmail and Outlook have built-in tools for this - and Gmail has a hidden shortcut most people have never seen.

In Gmail

Go to Gmail in your browser. In the URL bar, replace everything after /mail/ with #sub:

https://mail.google.com/mail/#sub

It pulls every subscription into one place.
Work through them with your keep list in mind. Anything not on it - unsubscribe.

In Outlook

Search “unsubscribe” or use Settings → Mail → Subscriptions.

Same idea. Find them. Remove them.

Step 2 — Archive the backlog

You are not sorting this. You are removing it from view.

Archive everything older than a year. Not delete - archive. The box moves to the basement, not the dumpster.

In Gmail

Search:
before:2024/1/1 is:unread
Select all → Archive

Done.

In Outlook

Filter by date → select older emails → Archive

That’s it. It’s still there if you need it. You just don’t have to look at it.

Step 3 — Make this manageable

If your inbox is truly out of control, one sitting won’t do it. Block twenty minutes a day until it’s clear. Morning coffee, lunch break, before you close your laptop. Wherever it actually fits.

Once it’s clear: two email windows a day is plenty. Morning to clear overnight. Late afternoon before you close out. The rest of the time, close the tab. Almost everything can wait.

What I didn’t do

I didn’t create labels. Didn’t set up filters. Didn’t sort anything before archiving it. Didn’t spend an afternoon deciding what sparked joy one email at a time - that’s Marie Kondo, and this is not that. This is ruthless.

And I didn’t chase inbox zero - because inbox zero is not the point. The point is being able to see what’s actually in your closet. A manageable inbox isn’t empty. It’s just one where you can find the things that matter.

The labels come next post. The filters come after that. The part where AI handles most of this without you lifting a finger - that’s where this is all going. This step is just the purge. Clear the space first. Everything else has somewhere to land once you do.

Here’s what I want you to picture: you open your inbox and the noise is already gone. The things that actually need you are right there. The rest is handled - sorted, labeled, dealt with - before you ever see it. You didn’t do any of it that morning. It just ran.

That’s where this series ends. But we get there one step at a time. And you can’t organize a closet you can’t see into.

Clear it first. That’s what this is.

Less noise. Fewer things slipping through the cracks. A much better chance you actually see the birthday party invitation before the party happens.

That’s your little win - and it’s the foundation for everything that comes next.

Block the time.

The Daily Cleanout

Twenty minutes a day until the inbox is clear. Add it now before you close this tab.

Add to Google Calendar

The Weekly Evite Search

Until your inbox is clean enough to catch these automatically, run this search once a week so nothing slips through. Copy and paste directly into Gmail:

from:evite OR from:paperlesspost OR from:punchbowl OR subject:birthday

Block ten minutes every Sunday - or whatever day you do your weekly reset - and run it. Your kids will thank you.

Add to Google Calendar

The slightly smarter version

Skip the weekly search entirely by creating a Gmail filter that automatically labels these emails the moment they arrive. Go to Settings → See all settings → Filters and Blocked Addresses → Create a new filter. Paste the search string above into the From and Subject fields, click Continue, and apply a label called “Invitations.” They’ll be waiting for you - flagged, findable, impossible to miss.

Next post: the closet is clear. Now we organize it - labels, structure, and your first taste of AI actually doing something useful in your inbox. Once things are labeled, you can find anything in seconds.

I co-chaired an entire Colonial Day for my daughter’s fourth grade class. Full event. Costumes, activities, a full day of controlled chaos for 50 kids. I was reaching for a well-earned Aperol spritz when my phone buzzed.

What are we doing for the end-of-year party?

I put my phone down. I stared at the wall. I took a long sip.

This is May. This is always May. The calendar looks manageable until it suddenly isn’t - and by the time you realize it isn’t, you’re already behind on three things you didn’t know were on your list.

has been talking about this for years. If you don’t know her, fix that immediately - her Bright Method is the most practical system I’ve found for actually managing your life instead of just reacting to it. Kelly thinks of everything so you don’t have to. She even has a ‘Maycember’ prep list that maps out the entire season before it hits you. It’s brilliant. The problem is it requires you to sit down at your kitchen island with your computer and actually do the work.

I did the Bright Method program and I love Kelly’s system. I also have two kids, a full-time job, and approximately zero minutes of sit-down energy left by the time May rolls around.

So now I do an Offload. On a walk. Talking to my phone.

The Offload.

When the spiral hits, I open Claude and do a brain dump by interview. I start a chat and say: I need to get everything out of my head about the end of school year and summer. Ask me questions one at a time until we’ve found it all.

And it does. Slowly, systematically - school events, camp logistics, teacher gifts, travel windows, doctor appointments that need to happen before August, the things I’m doing versus the things I’m coordinating for someone else. It keeps going. Anything related to activity sign-ups? Any family travel that affects school pickup?

Twenty minutes later I have a list I never would have built alone. Not because I’m disorganized. Because the list lives in ten different places in my head and the only way to get it out is to be asked for it, piece by piece.

Then I say: Now turn this into calendar blocks.

It does. Actual events, with names and dates - already on my calendar. Done. I don’t then have to open my calendar and manually recreate what I just spent twenty minutes surfacing.

You can run The Offload in ChatGPT - and it will absolutely help you get everything out of your head. The reason I use Claude for this is the execution layer. When I say “now turn this into calendar blocks” and it just does it - that’s a Claude thing. ChatGPT will get you to the list. Getting from the list to the calendar takes a few more steps. Worth knowing before you start.

This works for anything you’re managing, by the way - not just kids and school years. A move, a work project, a trip, a renovation. If it lives in your head in pieces, The Offload will find it.

The Part Most People Skip

Most people use AI the way they use Google. Type something vague, get something generic, feel vaguely disappointed, move on. The tool isn’t the problem. The setup is.

If you’ve been reading The Little Wins for a while, you’ve already met my AI Stylist and my Weekend Concierge. Those aren’t just cute names - they’re a real way of thinking about how AI works best. You give it a role, you brief it properly, and then you have a conversation with it the way you would with an actual expert. Here’s the framework underneath all of it.

The Brief.

This is your context file - everything your expert needs to know about you before the conversation even starts. Think about the last time you sat down with a good stylist, a financial advisor, a trainer. You didn’t just show up and say “fix me.” You talked. She asked questions. She learned your life before she gave you a single recommendation.

AI is the same. In ChatGPT, your Brief lives in Project Instructions - background that loads into every conversation automatically. In Claude, it lives in a Markdown file inside a Project. Set it up once and every conversation after that starts already knowing you. Your concierge already knows you have two boys and zero interest in the Taylor Swift festival. Your stylist already knows you hate anything dry-clean only. The difference between a Brief and no Brief is the difference between a recommendation that fits your life and one that could belong to anyone.

The Ask.

This is how you start the conversation - but it’s just the start. Think about sitting in a chair at the salon. You don’t walk in and say “cut it” and then sit in silence. You talk. How much off? Are we doing layers or keeping it one length? You pull up a photo on your phone - is it the length you want or the texture? She makes the first cut, you look in the mirror, you give feedback. A little more here. Keep the length there. The whole conversation is the work. Your prompt is just how you open it. You don’t need a perfect prompt. You just need to start talking.

The Done.

This is what happens once you have clarity. Claude doesn’t just help you think - it executes. It builds my calendar blocks. It turns a list into a week I can actually live. That’s not a conversation anymore. That’s a second pair of hands.

The Set and Forget.

This is where it’s all heading - when your expert stops waiting to be asked and just handles it. The weekly family review that drafts itself. The Sunday night briefing that lands in your inbox before Monday hits. We’ll get there. But start here.

The two prompts. Copy them. Use them this week.

Prompt 1: The Offload. Use it now.

Prompt 2: The Brief. Use it once, save it forever.

When it’s done, paste it into your ChatGPT Project Instructions or your Claude Markdown file. Every conversation from here knows who you are before you type a word.

Run Prompt 2 first. Then come back to Prompt 1.

Less invisible load. Better summer visibility. More mental space for the things - and the Aperol spritzes - that actually matter.

That’s your little win.

Before you go - quick question.

I’m working on something to help with summer camp logistics and I want to make sure it actually solves the right problem. Pick the one that makes you want to cry the most. One click, that’s it.

Subscribe now

I paid my tax bill this week. I’m not going to tell you the number, but I will tell you there were tears. Which is embarrassing, because I knew it was coming, I had the money set aside, and I am - by most reasonable definitions - a person who has it together. And yet, there’s something about watching that number leave your account that no amount of preparation fully cushions. I always plan ahead - quarterly check-ins, income tracking, the whole thing. But "plan ahead" used to mean opening my calendar, creating four separate entries, writing the notes, setting the reminders - the kind of setup that takes twenty minutes and still somehow feels incomplete. This year I just asked. Five minutes, done, not touching it again until July. Same intention I've always had. A fraction of the effort.

Here’s what got assisted this week.

Helping Future Me

I stopped dreading next tax season before the tab was even closed

After paying the bill, I opened Claude CoWork - think of it as an automated tool that can actually do things on your computer, without you having to open a single app or create a single entry yourself - and asked it to schedule quarterly financial check-ins for the rest of 2026. Blocked on my Google calendar, labeled, done. Each one has a simple attached note: review income trends, confirm funds are set aside, adjust if needed. No elaborate system. No spreadsheet I’ll abandon by June. The whole point was to get it out of my head and onto a calendar before I talked myself out of it. I don’t have to remember to remember anymore.

Stop the Scroll

I shopped the Sephora sale like a professional. Because in another life, I was.

Here is something my friends know about me: I love makeup. Not in a “I wear a little mascara” way. In a “I spent the early 2000s popping lashes on everyone in the room and I still get asked to do friends’ makeup before events” way. In a “I will watch a 45 minute tutorial about a highlighting technique I will never use just because it’s beautiful” way. I love a glow. I love a product. And every year I tell myself I’m not going to spend four days watching strangers on the internet tell me what to buy - and every year I absolutely do. This year I didn’t. I took photos of what I actually use and asked Claude to help me figure out what was worth upgrading and what I was just buying out of habit.

For hair I flagged something I’m navigating now: thinning. New territory. I asked it to factor that in specifically. What came back wasn’t a trending product roundup - it was a focused list built around what I already had, what had changed, and what would actually move the needle. I still spent money. I’m not a different person. But I spent it on the right things. For someone who has been known to spend three days researching a moisturizer, that is personal growth.

And just in case you’re curious - here’s what actually made the cart. No affiliate links. No agenda. Just in case you love a product as much as I do.

• LYS Beauty No Limits Cream Bronzer Stick — $24

• Makeup By Mario Ethereal Eyes Eyeshadow Palette — $69

• Color Wow Raise the Root Thicken & Lift Spray — $26

• Nécessaire The Scalp Serum — $48

• Armani Beauty Luminous Silk Foundation — $69 (a forever foundation - but a new formula)

• Sephora Collection Cream Contour Stick — $16 (proof that the best products don’t always cost the most)

Show It to Me

I helped a friend get the AI try-on to actually work

A friend had been trying to get a realistic AI try-on of a dress she found online. She’d been at it for a while - frustrated, ready to give up - and sent me a “why does this never work” message. I sat down with the AI agent myself, worked through the whole thing, figured out exactly where it was going wrong, and then recorded a walkthrough video showing her what to adjust and why. She watched it and texted back: “Wow — this is so detailed.” It worked on the next try. No starting over, no new apps.

BTW - I made that video using the screen record feature that’s already on your phone. If you didn’t know it was there: on iPhone, swipe down from the top right corner and tap the record button. On Android, it’s in your quick settings panel. That’s it. No app, no setup. You’re welcome.

If enough of you want a version of that walkthrough video for the newsletter, just reply and let me know.

Chat Therapy

I helped a friend think — without giving her my opinion

Another friend is weighing a job offer that would mean relocating. The kind of decision that’s impossible to think clearly about when there are too many feelings in the way. I didn’t offer my take (I’ve learned). Instead I sent her a prompt — and I want to be specific about this because the details matter. I didn't just say "ask Claude to help you decide." I gave it context, a coaching style I actually trust, and a clear set of instructions for how I wanted it to show up:

“I need help working through a big decision. I want you to approach this the way Brené Brown would - warm, curious, no judgment, focused on helping me understand what I actually value and what I might be afraid of. Please interview me one question at a time. Don't give advice. Don't offer opinions. Just ask me questions that help me get honest with myself. Start by asking me to describe the decision I'm facing.”

She came back an hour later having worked through more of it than any conversation with a friend would have gotten her to. AI didn’t make her decision. It just gave her space to hear herself think - without anyone else’s feelings in the room. Go ahead and steal that prompt. Swap in whoever your person is. It works for anything.

Wait, Explain That

The Masters moment that caught me off guard

I want to be clear: I have been to golf events. I follow along. I know how a leaderboard works. But I was watching the Masters this week and one player’s score looked so dramatically out of step with everyone else’s that I genuinely paused. Not a “what is golf” moment - more of a “wait, why does that number look like that right now” moment. The kind of thing where you know enough to know something looks off, but not quite enough to know why. Rather than let it quietly bother me for the rest of the round, I just asked: Why does this player have such a high score when everyone else is so low - is that normal at this stage of the tournament? Turns out he’d already finished his round. Everyone else was still on the course. The leaderboard looks completely different mid-round than it does when the dust settles, and once you know that, it’s obvious - but nobody explains it because everyone assumes you already know. A moment that would have quietly bothered me for the rest of the round - resolved in thirty seconds.

None of this is a breakthrough. A tax plan that took ten minutes. A shopping list I actually feel good about. Two friends who had a slightly better week. One less thing to quietly wonder about during a Sunday afternoon. Small, specific, done.

That’s The Assist.

P.S. I’m told that every newsletter about women and AI should include a meal planning tip. I’ve never used AI for meal planning and I don’t intend to start. If that’s why you’re here, I hope you’ll stay anyway - it gets more interesting.

We’re not those travelers.

We don’t plan restaurants and meals in advance.
We don’t lock in every museum.
Frankly, we almost never make reservations unless we absolutely have to.

Part of it is logistical - life is busy and planning takes time.
But part of it is intentional.

We like to wander.

We like waking up and deciding what the day looks like over coffee.
We like saying yes to something unexpected.
We like not feeling locked in.

We also travel like this… with kids.

It feels like it probably shouldn’t work and honestly, it sometimes hasn’t.

But this is how we like to travel - flexible, a little unstructured, and determined to figure it out as we go.

It just used to come with more friction.

We’ve had multiple trips where we’ve:

• frantically searched for a bathroom or convenience store in an Italian ghost town when everything is closed for afternoon “riposo”

• eaten canned tuna for dinner because we didn’t have the energy to find a local restaurant

• or realized too late that something we should have planned ahead for was completely booked

This time, everything just felt easier.

Same style, same level of “let’s just figure it out”, but with fewer hiccups.

The difference was simple:

I had ChatGPT in my back pocket the entire time.

Not in a big, over-the-top way.
Just in a quiet, incredibly useful, this is making my life easier kind of way.

Here’s how I actually used it over the past week.

1. Things I Normally Would Have Given Up On

Let’s start small - but very real.

The combination washer/dryer.

If you’ve ever tried to use one in Europe, you know. It looks simple until suddenly you’re staring at symbols, multiple compartments, and a cycle that somehow takes 5X as long as it would in the US.

In the past, I would have:

• avoided doing laundry entirely

• or guessed and hoped for the best

Instead, I took a picture of the appliance and asked ChatGPT to explain to me how to use it.

That is literally all I asked at first. And then, I just kept taking pictures and asking questions, because I couldn’t believe it was actually working.

Things like:
“Now I see this - is that right?”
“I hear water, but I don’t want to run a wash cycle…”
“Am I ever going to be able to get my clothes out? It says it’s locked for 5 hours?”

And it just walked me through it!

2. Handling a Flight Cancellation Without Spiraling

A week before leaving, our Aer Lingus flight to Italy was canceled and automatically rebooked - routing us through Heathrow with a massive layover and getting us in almost a full day later than originally planned.

We had a family event scheduled for Saturday afternoon and missing it wasn’t an option.

Normally, this is when I panic and mentally shut down.

Instead, I opened ChatGPT and said:

“Our flight was canceled and rebooked with a huge delay - what can we actually do here?”

What came back was surprisingly clear: we had real options, and more importantly, we had leverage.

I learned that under EU law, we weren’t stuck with what they gave us - we were entitled to be rerouted at the earliest opportunity, even on a different airline, and in comparable conditions.

That changed the entire conversation.

From there, I had ChatGPT help me find better options - flights that would actually get us into Italy in time, into either Rome or Naples. Chat gave me specific routes and flight numbers.

So instead of asking the airline “what are my options”, I went in with a specific goal in mind - getting us into Rome or Naples on Friday or Saturday morning.

And then I asked the most helpful question, “What should I actually say when I call them?”

It gave me a simple script. Not formal, just direct:

“I’d like to be rerouted under EU261 at the earliest opportunity. The current itinerary adds a full-day delay. Can you move us to one of these earlier options?” And then I gave them the specific routes and flight numbers that Chat provided to me.

That was it.

When I got on the phone, I wasn’t guessing.

I knew:

• what I needed (Saturday arrival)

• what options existed

• and what I could reasonably push for

It didn’t make the situation go away.

But it turned what would normally feel scattered and stressful into something much more manageable.

And after we got everything sorted, I went back to it one more time—this time to figure out how to actually file for compensation.

I asked it to walk me through the EU261 claim process and help me fill out the form correctly - which I probably wouldn’t have done otherwise.

A week before we left, this trip was almost canceled.

Instead, it got rebooked, we made it to everything we needed to - and there’s a real chance we may actually get a meaningful amount of money back (up to €600 per person).

All because I asked a few questions I wouldn’t have known to ask before.

3. Packing Without the Mental Load (With a Little Help From My AI Stylist)

I hate packing.

I used to try to make it easier by keeping very detailed packing lists - honestly, embarrassingly detailed.

I had different versions for different types of trips, and I would share them via Google Drive with anyone who asked.

It helped… a little.

But it still felt like a lot of mental work:

• What am I forgetting?

• Am I overpacking?

• Do these things actually go together?

Now, I just use ChatGPT.

I’ll start with something like:

“Help me plan outfits for a weeklong trip to Italy where I need something appropriate for Easter Mass, a nice dinner out, but mostly days of walking, wandering, and traveling. I want to look put-together but not like a tourist. Keep it minimal - I don’t want to check anything - and limit me to two pairs of shoes.”

But then I keep going.

I’ll layer in more context as I think of it:

“The weather looks like this…”
“We’ll be in the mountains for a few nights and it’s cooler there”
“I definitely want to wear this sweater - can you build around that?”

And then I’ll literally try things on in front of a mirror, take pictures and ask:

“Do these pieces work together?”
“Am I overpacking?”
“What would you cut?”

It becomes much more of a back-and-forth with my “AI Stylist” than a one-time plan. I wrote more about how I set this up in a previous post.

Instead of overthinking every outfit, I’ll have it help me pick a general look or color palette for the trip - something cohesive so everything mixes and matches without needing ten options.

And then I refine as I go.

I even asked it how to tie a scarf like a European - which felt like a level of preparation my old packing spreadsheets never quite reached.

It didn’t pack my suitcase for me.

But it took the thinking off my plate.

4. Having a Calm, In-the-Moment Travel Brain With Me

I didn’t just use ChatGPT before the trip - I used it constantly during the trip.

For all the random, real-life questions that come up when you’re actually in the thick of it:

• “What does this street sign mean?”

• “Can we park here?”

• “When do we need to leave for the airport in order to return the rental car and make our flight at 11:00?”

• “Why has it taken us an hour and a half to drive one block?”

• “Is this underground cave tour going to freak my kids out?”

• “Can I get last-minute tickets to see the Veiled Christ?” Spoiler alert: no.

But also for something I didn’t expect to use it for:

Quick, on-the-fly context.

The kind you’d normally skip because you don’t have time to read an article or look something up properly.

Like standing outside the tomb of San Gennaro and asking:

“Can you give me a quick, kid-friendly version of who San Gennaro is before we go in?”

And getting back exactly what I needed:

• who he was

• why he matters in Naples

• and the detail about his blood - how it’s said to liquefy a few times a year and people still gather to witness it

Which, as you can imagine, immediately got their attention.

It wasn’t a full history lesson.

It was just enough to make the moment more interesting and meaningful - for all of us.

That’s really how I used it most of the time.

Not for deep research.

Just for quick, clear, “tell me what I need to know right now” answers.

Some of these are things I probably could have figured out… eventually.

But I didn’t need to.

Each time, I got a quick answer, the context I was missing, and a lot less second-guessing

Because most of the time, I don’t need a deep dive.

I just want the 20% that makes the moment make sense.

And that’s exactly what it gave me - right when I needed it.

5. Making a Flexible Trip Actually Work

Here’s the reality of how we travel:

It’s fun.
It’s flexible.
And it depends on figuring things out as we go.

On our way down to Naples after visiting family, we wanted to stop somewhere.

We had already been to a lot of the obvious places - including Caserta.
We didn’t have a plan, and we weren’t about to spend time researching from the car.

So I asked, “We’ve already been to Caserta - what are good stops on the way to Naples that aren’t the main tourist spots?”

Within minutes, we had a few solid options.

This is basically what I’ve been training my “weekend concierge” to do—just applied in real time.

And we kept using it that way.

When we realized we hadn’t gotten tickets to the Veiled Christ, instead of going down a research rabbit hole or feeling like we missed something, I just asked:

“What should we do nearby that’s actually worth it?”

And again, we had a plan in minutes.

We’re still going to travel the same way - loosely planned, figuring it out as we go.
This just makes it a lot easier to do that well.

What ChatGPT Didn’t Do

It didn’t make the trip perfect.

We still missed things we didn’t plan ahead for.
Had meals that were just okay.
Changed plans constantly.

That’s just how we travel.

The Bigger Shift

What I love about using AI this way is how low-stakes it is.

I didn’t need to learn a system or do hours of research or get everything right ahead of time.

I just asked - and adjusted as we went.

And it worked.

Because the biggest thing I noticed on this trip wasn’t that everything went perfectly.

It was that we didn’t need to have everything figured out ahead of time for it to work.

We’ve traveled to Italy for years.
My husband speaks the language.
We know how to do this.

And we still relied on ChatGPT constantly.

Not to plan the trip.

Just to make it all work a little better as we went.

Less second-guessing.
Better decisions.
More ease.

I’ve been using ChatGPT like this more and more - travel, planning, even the random questions I used to just ignore.

If this sounds like how you travel (or how you wish you traveled), send it to a friend you’d take a trip like this with.

Share

And if you want more ideas like this - simple, real ways to use AI in everyday life - you can subscribe here.

Subscribe now

I have every reason to be wary of AI.

Of all the professions it has the potential to disrupt, and frankly, dismantle - legal is a clear frontrunner. AI can digest and synthesize complex documentation in minutes. That template I’ve been refining for 20 years? It can produce a solid version in seconds. The concise, TL;DR client email that’s been a hallmark of my practice? Done in under a minute.

When I first started experimenting, I felt it immediately - a creeping sense of irrelevance.

It was clear that AI could do much of what I do - faster, and if I’m honest, sometimes better. I felt threatened.

So under duress, I did what felt like the only rational move - I kept my enemy close.

And then, something else happened…

I started inviting Chat into everything.

Not the high-stakes, “this needs to be perfect” kind of work - the opposite.

Small, random, low-takes things that I normally wouldn’t have spent time on – because I didn’t feel fully prepared, or because it felt like too much effort to get started.

The kinds of things I used to try in my 20s, when I had more curiosity than time constraints, and far less concern about whether something was “worth it.”

Back then, I would sign up for things on a whim. A capoeira class. African drum dancing. A lecture on something I barely understood. Not because I had a plan – just because it sounded interesting.

Some of it stuck; most of it didn’t.

Somewhere along the way - between careers, kids, logistics, and the constant background noise of responsibility - that instinct dulled.

Everything started to feel like it needed a reason, a return, a clear outcome…

And if I didn’t feel fully prepared? I just didn’t start.

Over the past year, using Chat has done something I didn’t expect.

It has given me energy.

Not the fleeting kind that comes from a new tool, but a steady sense of momentum. The feeling that I can try things more easily, follow ideas a little further, and stay engaged in ways that used to feel just out of reach.

What changed wasn’t just speed - It was willingness.

I started more things. I hesitated less. I explored areas where I didn’t yet feel fluent.

Because I was no longer starting alone - and because it no longer felt like such a heavy lift just to begin.

But even more than that, it has changed how learning feels.

I don’t have to find a class or go to a conference.
I don’t have to buy a book.
I don’t have to leave the house.
I don’t have to commit to something before I even know if I’m interested.

I can sit cross-cross applesauce in my athleisure and just try.

For a few minutes - in the middle of my day or at 11:00 at night – without it turning into a whole thing.

And that has made me feel more engaged with the world around me than I have in a long time.

What’s surprised me most is how normal the use cases actually are from day to day.

I’m generally not doing anything particularly technical. I’m mostly just… talking to it.

And over time, a few patterns have emerged:

Sometimes it’s a thinking partner

When I’m stuck or overthinking something, I’ll write out what I’m considering and ask for a reaction. Or even better, I will ask it to interview me.

Even something as simple as, “What am I missing here?” is often enough to get me unstuck.

Sometimes it’s a way to learn on demand

I’ve used it to better understand news stories I only half-followed, to get background on topics I felt like I “should” understand, and to go deeper when something piqued my interest.

Instead of skimming, I can actually understand what’s going on.

And over time, that adds up.

Sometimes it’s a low-pressure teacher

I set up a simple Italian tutor so I can practice whenever I have a few minutes.

No schedule. No pressure. No feeling behind.

Just picking it up when I feel like it – and focusing on the subjects that I find most challenging.

Sometimes it’s a parenting backup

A math concept I don’t immediately remember how to explain (and frankly, probably never understood in the first place).

A request to create an outline for a school project so that it doesn’t quickly become my assignment.

An empathetic response when your daughter tells you that the mean girl was mean to her (again).

It’s like having an extra set of hands (and patience) when I need it.

Sometimes it’s the adult in the room

There are areas of life where I’ve always felt like I was… mostly keeping up.

Investing is one of them.

Recently, I opened my investment dashboard and realized I was looking at a lot of information I wasn’t fully using.

So I sent a screenshot of the Fidelity dashboard (with personal information obscured) and asked, “Can you walk me through how to read this like someone who actually understands what they’re doing?”

Sometimes it’s a curator

If you’ve read my AI Stylist or Weekend Concierge posts, you’ve seen the output of this.

Travel ideas. Reading lists. Weekend plans.

Those posts are essentially curated - pulling together options that fit a specific mood, moment, or need.

The difference now is that instead of starting from scratch - or falling down a rabbit hole of tabs - I can start with a point of view and build from there.

Sometimes it helps me see things

I frequently use it to create rough visual mockups of architectural or design decisions.

Questions like, “Can you create a visual mock-up of our front yard with these two fence options?” or “Can you show me what our dining room would look like with this lighting fixture?”

Often, it’s a way to simply pursue curiosity

The small questions I might have dropped before - because they felt too minor or too time-consuming - are now easier to explore.

Most of them go nowhere.

A few of them turn into something useful.

But all of them keep me engaged.

And somewhere along the way, something else has started to happen. I’ve started building things. Not in a technical, “I know how to code” way. More like describing what I want, adjusting it, shaping it - until something useful exists where there was nothing before.

A small tool. A system. A starting point I can reuse.

It’s not something I would have ever thought of myself as doing. But it has quietly made me feel more capable than I have in a long time.

If you haven’t really used tools like this yet, the hardest part is just starting.

So here’s a very simple way to try it:

Pick something from your real life.

A question you have. Something your child needs help with. A topic you’re curious about. A decision you’re thinking through.

Ask it the way you would ask a person.

No special language needed. Just write what you’re wondering.

Then ask one more question:

“Can you simplify that?”
“Can you give me a few options?”
“Can you make that more specific to me?”

Use what’s helpful. Ignore what isn’t.

Try it again the next day.

The value isn’t in one interaction - it’s in building the habit.

Give yourself 15 minutes a day.

If you do this a few times, something subtle happens. You start to trust that you can begin without having everything figured out. You rely less on perfect preparation and more on iteration. And over time, that changes how you work - and how you approach new things.

I still think it’s reasonable to be wary of AI - but I also think we’re missing something if that’s the only lens we use.

Because for me, the most meaningful impact hasn’t been about replacing work.

It’s been about restoring something I didn’t realize I had lost.

Energy. Curiosity. A willingness to try.

And if there’s one thing I didn’t expect, it’s this - I’ve come out of this feeling more capable, more engaged, and, in a very real way, a little smarter than I was before.

And that may actually be the biggest win of all.

Everyone has a friend that is really, really good at having fun.

She never misses the fall festival. She somehow buys concert tickets before they sell out. She notices the random teacher workday early enough to turn it into a long weekend away. Her family always seems to be doing something memorable, charming, or just slightly more interesting than everyone else.

Meanwhile, the rest of us have the same thought – typically on Friday afternoon: we should do something fun this weekend.

And then we order pizza and watch Netflix.

It’s not because we don’t want to do something – it’s because we’re too tired to plan, much less do.

After a full week of work, school logistics, emails, meals, scheduling, and everything else, even fun feel likes work. It becomes one more thing to research, one more decision to make, one more task to manage.

By the time the weekend arrives, most of us don’t need more options - we need fewer decisions.

All week long, we’re carrying the mental load—planning, coordinating, remembering, anticipating.

By Friday, even something enjoyable can feel like one more thing to figure out.

Meet my Weekend Concierge

That’s where my “Weekend Concierge” comes in. This is one of my favorite examples of what an AI Project can do when you stop using ChatGPT like a one-off question box and start using it like a real expert you can train over time.

My Weekend Concierge is not just a list generator or a better search engine. It’s a chief of staff that actually knows me and my family:

• It knows that I generally hate a musical – with Les Mis and &Juliet as limited exceptions

• It knows that when the weather turns nice, I am not looking to run a 10K – I just want to sit outside somewhere pleasant and have a cocktail.

• It knows that I love Americana but am also a closeted Europhile.

• It knows that my family generally loves a charming, slightly under-the-radar plan - an antiques festival, a spring farm day, Embassy Day in DC - something that feels a little special but not overwhelming.

And the real magic is this – I do not have to start from scratch every single time.

Instead of opening my laptop on a Friday afternoon and trying to pull together a weekend of fun, I have a dedicated expert that already understands the assignment.

All week long, while I’m doing everything else—working, answering emails, packing lunches, coordinating carpools—it’s quietly paying attention.

It doesn’t interrupt me. It doesn’t ask me to make decisions in the middle of a busy Tuesday.

It just prepares, so that by the time I think, we should do something fun this weekend, the planning is already done.

It might notice that there is an antiques market in Purcellville this weekend.

That one of our favorite farms has just opened a seasonal event.

That there is an embassy open house we would genuinely enjoy, not just the busiest one everyone else is going to.

That Sunday afternoon has the best weather window for something low-key outside after Italian school.

The Weekend Concierge is not just about finding activities – it is about reducing mental load. It is about removing the invisible work required to make modern family life feel enjoyable.

It is about planning the fun before I am too tired to care.

Set Up the Project

Under the hood, this is actually pretty simple.

I built this as a standing Project inside ChatGPT. Think of a Project as a dedicated space for a specific role or area of life. In this case, the role is personal concierge and lifestyle planner. Inside that Project, I gave ChatGPT background on my family, my preferences, my location, and the way I want ideas delivered.

In other words, I did not just ask, what should we do this weekend?

I built the context first.

Here is a simplified version of my Project Instructions:

You are my personal concierge and trusted lifestyle advisor. I’m a busy mom and the CEO of my family and home. I need your help making thoughtful, aligned decisions across all areas of life that require my input — especially activity planning.

We live in the metropolitan Washington DC area – specifically, in Falls Church, Virginia.

When suggesting activities, please keep these preferences in mind:

• Friday nights are lower energy evenings. We generally prefer adult-only activities on Thursday, Saturday, or Sunday nights.

• We love antique festivals, farm festivals, equestrian events (particularly show jumping and three day eventing), outdoor concerts.

• We will attend professional sports events (especially golf or tennis), but they are not a top priority.

• We love all things European – especially Italian.

• Favorite special events in the past have included – the Brimfield Antiques Festival in Brimfield MA, Lobster Festival in Rockport, MA, the Mid Atlantic Antique & Garden Festival spring show in Poolesville, Maryland, Embassy Days in DC.

I often carry the mental load of our household, so your job is to reduce my overwhelm and proactively bring clarity. Think of things I may not be thinking of. Recommend ideas, ask smart clarifying questions, and always organize your suggestions clearly (e.g., in steps, options, or short checklists).

Be polished but warm in tone. Help me feel more on top of things and confident in my decisions. When appropriate, gently remind me of my previously stated goals or next steps. Remember activities that we liked and those that we disliked. Adjust your recommendations accordingly.

Once this is set up, weekends start to feel different.

Not busier, not over planned, just easier

You’re not staring at a blank slate on Friday afternoon, trying to summon the energy to figure something out. You already have a few thoughtful options in front of you—things that actually fit your life, your family, and your mood.

And over time, it goes beyond just the weekend.

You start to have a sense of what is coming up this month.
What might be worth planning a little further ahead this season.
What you don’t want to miss entirely.

Some weekends you’ll still order pizza and watch Netflix. (Highly recommend Vladimir, by the way)

But other weekends, you’ll find yourself at the Mid-Atlantic Antique & Garden Festival, or sitting outside in the Arboretum, or doing something that feels just a little more memorable—without it taking extra effort to get there.

That’s the shift.

Not more pressure to do more.

Just a little more intention, with a lot less mental load.

I recorded a few quick videos to show exactly how I set this up and how I use it in real life—because once you see it, it becomes very doable.

And if this gives you one less decision to make this week, or one more plan that actually happens?

That’s a little win.

For years, Instagram served as my stylist. When I didn’t know what to wear, I scrolled. I loved the endless stream of inspiration. It reminded me of my younger days – afternoons spent mindlessly flipping through the pages of YM and Seventeen - and later, InStyle and Cosmo.

But unlike those magazines, Instagram never runs out of pages.

What used to be a quick burst of inspiration started turning into the kind of scrolling where you look up twenty minutes later wondering how you got from “summer outfit ideas” to watching someone reorganize their band-aid bin.

At some point I realized I didn’t actually want to spend that much time on Instagram, but it had quietly become the place I went whenever I needed outfit inspiration.

The problem is that Instagram isn’t really a stylist.

It’s an endless stream of inspiration with no organization, no memory, and absolutely no awareness that you already own the same shirt dress in three colorways. And yet somehow, you’re still considering buying another one…

Like the magazines of our youth, the algorithm is excellent at inspiration — but terrible at actually helping you to get dressed.

That’s when I had a small realization.

Instead of letting Instagram style my outfits, I could build my own stylist.

So I opened ChatGPT and started experimenting.

My idea was simple: what if I could teach it my style the same way I might brief a real stylist?

Instead of scrolling through hundreds of outfits hoping something would stick, I could just ask questions.

How should I style this dress for a luncheon?
What shoes would work with this pants?
What should I pack for a three-day trip to Boston for a conference in February?

It turns out ChatGPT is surprisingly good at this.

You just have to start by teaching it your taste.

I gave it a few Instagram accounts whose style I love and asked it to analyze the aesthetic — the colors, silhouettes, patterns, and overall vibe that showed up again and again.

Then I asked it to act as my personal stylist using that style as a guide. And because styling advice is rarely one size fits all, I also told it that I’m only 5’1”.

Now when I’m standing in my closet wondering what to wear, instead of opening Instagram, I open ChatGPT.

Sometimes I take a photo of myself in an outfit and ask how to style it.
Sometimes I ask it to create a few outfit ideas using pieces I already own.
Sometimes I even ask it to source new items for the season.

Here’s exactly how to set it up:

1. Create the Project

First, create a dedicated project in ChatGPT.

If you haven’t used Projects yet, think of them as separate workspaces where ChatGPT can remember context and function like an expert.

You wouldn’t invite your dentist, your OBGYN, your therapist, and your stylist to the same dinner party and start firing off questions - each of those people works best when they’re operating in their own domain.

Your stylist needs to think about clothes.
Your therapist needs to think about feelings.
Your dentist needs to think about your teeth.

ChatGPT works the same way.

When you create a project, you’re essentially creating a private office space for your expert - in this case, your stylist.

This keeps all of your style questions, prompts, and conversations in one place.

2. Prep for Your Styling Session

Next, think about what you would tell a real stylist if you were sitting down together.

Stylists don’t magically know what works for you. They need context.

Start by jotting down a few basics.

1. Where do you live?

Style advice for someone in Florida looks very different from someone in Connecticut. Climate matters. Seasons matter. Culture matters.

2. Think about your size and proportions.

For example, I always include that I’m 5’1” because proportions make a big difference for me in styling outfits. You can also include your typical clothing size, and whether you have certain attributes (e.g., broad shoulders, short-waisted, etc.) or dislikes (e.g., I generally hate a cap sleeve).

3. Gather inspiration.

Open Instagram and pull together about 20 accounts whose style you consistently like. These might be designers, boutiques, influencers, or brands. You can collect them manually, or if you’re feeling ambitious, download your “Following” list from Instagram’s Privacy Center and pull inspiration from there.

The goal is simple: give ChatGPT clear instructions and a visual sense of your aesthetic.

3. Feed the Context

Paste the prompt (i.e., the context) into the Project Instructions.

Here is a simplified version of my personal styling prompt:

You are my personal stylist and aesthetic advisor.

My style is classic, tailored, polished, and slightly traditional. I like clothing that feels refined, timeless, functional, and quietly elevated — think Tuckernuck with a bit of Heather Chadduck sensibility.

When referencing my style, use these Instagram accounts as aesthetic anchors where helpful:

@m.g._style, @meganstokes, @thesoutherngail, @maddiepriorhullstyling, @heatherchadduck, @tuckernuck, @habituallychic, @kellymccoyd, @sara_ferguson, @things.i.bought.and.liked

I am 5’1” with a square/athletic build, so always keep proportion, scale, fit, and flattering silhouette in mind.

I am an attorney working in tech. When suggesting outfits, balance professionalism with a slightly more modern, stylish edge than a typical attorney wardrobe. Think tailored blazers, polished separates, elevated basics, chic flats, and clean sneakers.

I live in the Washington DC metro area, so keep four-season dressing and practicality in mind.

Use a polished, warm, and discerning tone. Be honest, specific, and constructive. Help me feel confident, pulled together, and clear in my decisions.

When I share a photo, give thoughtful, constructive feedback on fit, silhouette, styling, and what would make the look more flattering, polished, and cohesive. Always keep proportion, hem length, waist definition, scale, and overall silhouette in mind. I benefit from styling that creates polish and elongation without feeling fussy or overly delicate.

You are my personal stylist, wardrobe advisor, and aesthetic sounding board.

Your job is to help me get dressed, refine my wardrobe, evaluate outfits, recommend purchases, and strengthen my overall personal style in a way that feels polished, timeless, and fully aligned with who I am.

4. Let Your Stylist Style

Now the fun part - instead of scrolling Instagram for outfit inspiration, go to your stylist!

Here are a few specific questions that I’ve asked my stylist over the past month:

1. This outfit is not working. Why? What does it need?

2. I need a cocktail dress for a “Boots and Bling” themed event. I do not want to wear a “Western” themed outfit – instead, I’d like to embrace a “Dallas” vibe. Ideas?

3. Can I style these brown leather pants for a conference that skews tech – meaning that most of the men will be in elevated golf tech and trendy sneakers and there will be very few women on site?

4. I feel like I am always struggling to find the right spring travel jacket so that I can deal with cold temperatures and potentially wet weather but also look stylish for both sightseeing and dinner out. Do you have any ideas?

5. Try This Tonight

Create a “Personal Stylist” project in ChatGPT and add three things:

• your height
• your city or climate
• five Instagram accounts whose style you love

Then ask one question:

“Tomorrow, I am going to [Insert]. Give me three outfit ideas please.”

You might be surprised how good the answer is.

The Little Win

The funny thing is that the Stylist Project hasn’t made me think about clothes more. It’s helped me think about them less.

Instead of scrolling Instagram for twenty minutes hoping inspiration will strike, I can ask a quick question and get a few thoughtful suggestions based on my style, my height, and the clothes I already own.

It removes a surprising amount of friction from getting dressed.

And it means I can close the app and move on with my day.

That’s really the point of all of this.

Not using AI for the sake of using AI — but using it to solve small, everyday problems.

So you can spend less time scrolling; less time staring into your closet – and ideally, a little less time convincing yourself that you need another shirt dress or striped sweater.

A little win, but a very good one.

If you try this, I’d love to hear how it goes — or what prompts you come up with. The best little wins are usually the ones we share.