I Let AI Run My Mornings For a Week - Here's What Happened

I run an AI newsletter. I talk about AI constantly. And until last week, I was still making my to-do list by hand.

There's something more than a little ironic about that — but here's the truth: I'd been using AI for work. For writing, for research, for content drafts — all the professional stuff. The things I could justify. The things that felt safe to optimize.

My mornings though? Those were mine. Coffee, chaos, kids, calendar. A system that sort of worked and definitely wasn't pretty. It hadn't occurred to me to touch it.

Then I decided to try.

Not because I thought it would be revolutionary. Honestly, I wasn't sure it would help at all. But I write about this stuff. I tell people to test it in their real lives. It felt like it was time to stop exempting myself.

So I did. For one week. Every morning. And I'm going to tell you exactly what happened — including the parts that didn't work.

FIRST, WHAT MY MORNINGS ACTUALLY LOOK LIKE

Let me be clear: my mornings are not peaceful. There is no candle. There is no journaling in a quiet corner. There is a household that wakes up at full volume. Things that need to be signed, packed, found, remembered. A brain already running three tabs before I've finished half a cup of coffee. And somewhere in the middle of all of that, I'm also a founder with a newsletter to run and a business to build.

My mornings vary wildly. Some days I have an hour before things get loud. Other days I'm already behind before my feet hit the floor. There is no "typical morning" — which, I learned, matters enormously when you're trying to build any kind of AI routine.

I wanted to see what AI could touch in a morning, so I tested it on four things:
• Brain dump / journaling substitute
• Planning and prioritizing my day
• Drafting content ideas
• Handling messages and email

Here's what I found. Honestly.

This one I didn't see coming.

My old version of "journaling" was scattered notes on my phone — half-finished thoughts, things I didn't want to forget, worries dressed up as to-dos. It wasn't therapeutic. It was just noise in a different container.

I started doing this instead: opening Claude and typing whatever was actually in my head. Unfiltered. Sometimes it was about the newsletter. Sometimes it was something bothering me. Sometimes it was just the chaotic inventory of everything I needed to hold that day.

"Help me sort through this. What's important right now?"

Something about seeing it organized back to me — without judgment, without suggestions I didn't ask for — helped me think. It didn't give me answers. It gave me structure. And that was enough to move forward.

It felt less like using a tool and more like clearing a table before you start cooking.

I kept this one. It’s now part of how I start my day whenever I have even ten minutes.

I started giving AI my task list and asking it to help me prioritize.
The first couple of days, the output was fine. Logical. Reasonable. Completely missing the point.

Because AI doesn't know that Tuesday is harder than Monday for me. It doesn't know that one task on my list sounds small but will drain me completely. It doesn't know that certain kinds of work need to happen before noon or they don't happen at all.

This is where most people get AI wrong: they give it tasks. Not constraints.

Once I started giving it context instead of just a list — the output got genuinely useful. It would notice things I was burying. When I prompted it to, it would ask me whether I was putting the hard thing last because it was lower priority — or because I was avoiding it. That question, more than once, was uncomfortably accurate.

The catch: it only works if you show up like it knows you. It doesn't, not without you filling in the gaps. You're not handing your day off. You're thinking out loud with something that's very good at listening.

CONTENT DRAFTING: MY COMFORT ZONE — AND THAT WAS THE PROBLEM

This was the one I expected to nail. I do this every day.

But in the mornings, something was off. I was trying to use it before I was fully awake, before I knew what I actually wanted to say. And the output reflected that — fine, generic, the kind of thing that sounds like AI wrote it because, in that moment, I wasn't giving it enough of me.

AI can't draft its way to clarity you haven't reached.

Here's what I learned: AI for content creation needs you to show up with a point of view first. Even a rough one. Even half a thought. "I want to write about X and I think the angle is Y, even though I'm not sure yet" produces something entirely different than "write me content ideas about X."

The mornings when I'd already had a thought — something that nagged at me overnight, something I'd noticed — and handed that to AI to expand? Those were good. Really good.

The mornings I was hoping AI would hand me the thought? It couldn't. That part is still mine.

AI didn't fail here. I did. I was asking it for clarity I hadn't reached yet.

EMAIL AND MESSAGES: THE WILDCARD

Some mornings this saved me twenty minutes. Some mornings it added ten.

The saving-time version: paste in a message I'd been avoiding, add a little context — "this person is a potential collaborator, I'm interested but not ready to commit, warm but not overpromising" — and get back a draft that's 80% there. A few tweaks, sent. Done.

The adding-time version: messages that needed nuance I hadn't sorted out yet. I'd keep reprompting, keep adjusting, keep reading drafts that weren't quite right — and eventually realize the problem wasn't the AI. It was that I hadn't decided how I felt. You can't outsource emotional clarity you haven't found with the hopes that AI will process that for you.

I said it before (in this article, in fact) and I’ll say it again: AI can't draft its way to clarity you haven't reached.

THE PATTERN I DIDN'T EXPECT TO FIND

After a week, something became obvious:

AI didn't create clarity for me. It amplified the clarity I already had — and
exposed where I didn't.

Brain dump worked because the clarity existed, just scattered. Day planning worked once I added context. Content drafting failed without a POV. Email failed without emotional resolution.

The pattern held every single time.

The most useful part of this experiment wasn't speed. It was friction. AI
made it obvious where I was unclear, where I was avoiding something,
where I hadn't made a decision yet.

That's not a productivity tool. That's a mirror.

THE HONEST ENDING

A week in, I don't have a routine. I have a better understanding of where AI actually fits in my morning — and where it doesn't.

The brain dump stayed. The day planning stayed, with the condition that I show up with
context. Content drafting is still case by case — sometimes it's exactly right, sometimes I need to wait until I have more of a thought to give it. Email is situational.
None of this is a system - yet. It's a working relationship still finding its rhythm.

But here's what I keep coming back to:
AI doesn't replace thinking. It raises the cost of not thinking clearly.
That's not something you hear often.

If you come to this process foggy — unclear on your priorities, unresolved in your emotions, undecided on your point of view — AI doesn't smooth that over. It magnifies it. Which means the real ROI isn't in the tool. It's in the clarity you bring to it.
A tool that gets more useful the more clearly you know yourself.

Which, now that I say it out loud, might be the point all along.

IF YOU WANT TO TRY THIS TOMORROW

Here's the structure I use for the brain dump:
Here's what's happening: [what's on your plate]
Here's what's on my mind: [what you're carrying]
Here's what feels unclear: [what hasn't resolved yet]
Help me sort what actually matters.

Start there. See what comes back.
And if you're experimenting with AI in your own mornings, or anywhere in your life, I want to hear about it. What's working? What flopped? Reply if you’re subscribed or and get in touch. We’d love to hear from you!

The most interesting stuff always comes from real people doing real things.