I tried waking up at 5 a.m. When that meant too little sleep, I could manage a few good hours, and then I crashed. The rest of the day was a struggle.
Brooks naturally wakes early. I’m no sleep doctor, and I won’t turn his schedule or mine into a credential. Build a good start to your day around your sleep needs.
What actually sat between me and meaningful work was the preparation. Sorting email. Checking the calendar. Reviewing meeting history. Deciding where to start. I was doing all of that before I had touched anything that mattered. Now an agent assembles the context, and I begin the day with a direction.
Inbox triage can be the only reason for an early start. Hand that sorting to an agent and the administrative reason for the early alarm can disappear. A later start tells me that reason was gone. It does not tell me how much additional sleep came with it.
Hand the night-before plan to an agent
I still plan the night before. I used to do the whole pass myself. At 9 p.m. the repetitive part now runs as an AI workflow.
The agent reviews unfinished tasks. It inspects tomorrow’s calendar. It compares both against my goals for the next 30 days. From that, it proposes one thing to do first the next morning, then secondary options.
I kept the benefit of planning ahead. The preparation is what I delegated. The agent recommends. I decide.
Already doing some version of this by hand? Turn that evening checklist into a repeatable instruction. Give it the same inputs: unfinished tasks, tomorrow’s calendar, and your goals for the next 30 days. Ask for a single first priority and a short list of backups. Read the proposal before you adopt it.

Replace the inbox scan with a brief
Do you open email to figure out what your day is?
I stopped using the morning inbox scan as my launcher. An agent reviews email, calendar commitments, and relevant meeting history before I start, then surfaces the day’s priorities and the messages that need attention. I want a clear starting point and a short reply list. That’s the brief.
This is a workflow you configure. Connect only information sources you have approved. While you are still establishing whether you can trust it, check the brief against the originals.
One question before you start
AI can prepare my priorities. I still choose them.
Brooks warns that agents can gravitate toward tasks with due dates and overlook important work that is not urgent. A clean brief can still be aimed at the nearest deadline.
I let the agent assemble the options. Then I check the proposed first task. What important, non-urgent work is missing? If the top item only clears a deadline, I start with the work that advances a meaningful goal.
Flip the script
I used to ask for a personalized workflow and hope a generic prompt captured how I work. Now I flip the script. The agent asks me four to six targeted questions before it proposes a solution. Those questions reveal the context it was missing.
Paste this in before you ask for the brief:
“Before designing my morning brief, ask me four to six questions about my goals, schedule, information sources, and notification preferences.”
Answer them. Then look at what it proposes.
Two delivery windows
My updates arrive in two windows, and the rest of the morning stays quiet. A short morning brief carries priorities and essential information. Around 11 a.m. I get the fuller pass: email replies, follow-ups, and relationship news. The agent gathers in the background. My morning stays available for focused work.
11 a.m. fits me. Put your fuller delivery after your first focused block. Collection can run whenever. Interruption should wait.
The later window is where relationship news shows up. My agent reviews public updates about people I know and includes the relevant bits. A personal note, if one is warranted, is one I write and send myself. Monitoring on one side. The human note on mine.
Want to try a smaller version? Use public sources. Keep the pilot small. Decide yourself whether you respond and how.
Write the permissions down
My line is simple. Preparation can be automated. Actions that affect money or reputation require a human sign-off.
The agent can read relevant information and draft communications. Messages that go out under my name are ones I have personally reviewed. Finances stay read-only. Reputation-sensitive work stays read-only. Writes wait for manual approval. I send outbound messages myself, on purpose. I want the calmer morning and I want the final say on what I commit to.
Read access. List the email, calendar, and meeting history it may see. Mark finances and reputation-sensitive sources as read-only.
Drafting. Let it prepare priorities and draft replies. A draft still belongs to you.
Your approval. Money, commitments, and anything sent under your name wait until you say so.
Here’s the only move I’d make tonight. Identify one recurring morning preparation task. Just one. The email triage, the calendar check, the meeting-history review, or the decision about where to start. Hand that task to an agent with a clear instruction, and in the morning ask what important, non-urgent work it left out.
Recommended for you
Want the full system? 25X is the flagship productivity system we teach.
Explore 25X →