The First Hour Used to Be the Worst Hour
Here's what my mornings used to look like.
Open laptop. Check email. Respond to the thing that feels most urgent. Open Slack. Scroll. Open Notion. Stare at 40 open tasks with no obvious priority. Make coffee. Come back. Still no idea what to actually start.
By 10am I'd done a lot of responding and very little building. The day was already behind.
The problem wasn't laziness. It was that I had no system for converting the noise of an incoming day into a clear picture of what actually mattered. I was letting the most reactive things win every morning by default.
The AI morning stack fixed this. Not by eliminating the noise — the emails and Slack messages and open tasks are still there — but by compressing the "figure out what to do" process from an hour of scattered wandering to about 20 minutes of structured input.
Here's exactly what I do.
The Stack: Four Steps, 20 Minutes
The whole thing runs before I respond to anything. No email, no Slack, no notifications. That's the only rule that matters. If you start responding first, the morning is already someone else's agenda.
Step 1: Brain Dump with Claude (5 minutes)
Before I do anything else, I open a Claude conversation and dump everything that's in my head.
Not a to-do list. Not a structured plan. Just: what am I thinking about? What's nagging at me? What am I anxious about finishing? What do I remember I was supposed to do?
It comes out messy. That's fine. Claude's job in this step isn't to fix the mess — it's to reflect it back with structure.
My prompt:
"I'm going to do a brain dump of everything in my head this morning. Don't respond until I say 'done'. Then sort what I've said into: things I need to do today, things I need to decide, and things I'm worried about but can't act on right now."
I type for three to four minutes, say "done," and get back a clean three-column picture of my mental state. The "can't act on right now" column is important — it's the worry-holder. I look at it, acknowledge it, and then deliberately set it aside.
What this does is drain the cognitive load before work starts. If you don't externalise what's in your head, it sits there as background noise all day, consuming bandwidth you should be using for the thing in front of you.
Step 2: News and Context Digest (5 minutes)
I have a Perplexity Space set up for my industry. Every morning I run the same query:
"What's happened in [my industry] in the last 24 hours that I should know about?"
I don't read articles. I read the summary. I mark one or two things as worth following up on (these go into my Obsidian daily note, not my action list — they're reference, not tasks). Then I close it.
This step used to take 20–30 minutes of scrolling newsletters and news sites. Now it takes five. The AI doesn't replace the reading — it prioritises it. I only go deeper on things that actually matter.
If you're not in an industry where 24-hour news cycles matter, skip this step. Not everyone needs daily news digests. But if you do client work in a fast-moving field, walking into calls without knowing what happened overnight is a liability.
Step 3: Email Triage (7 minutes)
This is the one I was most resistant to and now can't work without.
I use Claude to process the first pass of my inbox. Not to write responses — to categorise. Each morning I copy the subject lines and first sentences of unread emails into a prompt:
"Here are my unread emails from this morning. Categorise each as: (A) needs a response today, (B) needs action but not today, (C) FYI only — no response needed, (D) can be ignored/deleted. Just the list, one line per email."
Thirty seconds later I have a triage list. I handle the A's now. Everything else gets batched into the afternoon processing block.
What this removes is the decision fatigue of opening every email and figuring out its urgency in the moment. That micro-decision, 40 times, costs more energy than people realise. Batching the classification into one AI-assisted pass at the start of the day means I'm spending that energy on the actual work instead.
I still write every response myself. Claude just tells me which ones need to happen now.
Step 4: Daily Schedule Block (3 minutes)
The last step is putting it together. I take the output from step 1 (today's tasks) and the A-list from step 3, and I ask Claude:
"Here are my tasks for today and the emails I need to respond to. I have roughly 6 hours of working time. Suggest a simple time-block schedule, putting the highest-focus work in the first three hours and batching the responses in the afternoon."
The output isn't a rigid schedule — it's a draft I adjust in my actual calendar. But having a starting point is dramatically faster than building the day from scratch each morning.
The important part is that step 4 only takes 3 minutes because steps 1–3 have already done the work of identifying what actually matters. By the time you're building the schedule, you're not making priority decisions anymore. You're just arranging things you've already decided.
How Long It Actually Takes vs. Before
The numbers above are my actual averages, not ideals. Some mornings take longer — if there's a crisis email in the inbox, step 3 expands. If I've had a busy week, the brain dump in step 1 runs longer.
But the floor is consistent. On a normal morning, 20 minutes and I'm at my desk with a clear first task and no lingering "what should I be doing?" anxiety.
The before numbers look small too, but that's deliberate. I'm not counting the 40 minutes I used to spend responding to emails before I started "real work." I'm counting the equivalent cognitive work: the time actively spent on orientation, not execution.
The Non-Negotiable
Do the morning stack before you check notifications.
This sounds simple. It isn't, because notifications feel urgent and the morning stack requires ignoring that feeling.
Here's what happens if you check first: your brain enters reactive mode before you've had a chance to assess what actually matters. The email from the difficult client feels like the most important thing. The Slack message about the minor operational issue feels like it needs handling now. You respond, and then you respond to the response, and by the time you remember you were supposed to do the brain dump it's 11am and the morning is gone.
The stack only works as the first thing. Not the second thing. The first.
I put my phone in another room until the stack is done. Overkill? Maybe. But I tried using willpower instead and it didn't work.
What to Start With
If you try to implement all four steps tomorrow morning, you'll build none of them.
Start with step 1. Just the brain dump. Do it every morning for a week. Notice whether starting work with a cleared head feels different to starting with a full one.
Once that's automatic — maybe two weeks in — add step 4. Build the daily schedule from your brain dump output.
Then add step 3 (email triage) once you've got a feel for how the structured morning changes your afternoon.
Step 2 (news digest) is optional and context-dependent. Add it if you need it. Skip it if you don't.
The goal isn't the stack. The goal is starting work every morning knowing exactly where to put your energy. The stack is just the fastest path I've found to get there.