Reading your daily brief is not the point. The point is that the work in it moves, including the work that should not be yours.
If your team has to interrupt you to find out what they should be doing, you are the constraint on their week, and a longer list of things needing you is a diagnosis rather than a to-do list. This article is the sequence that turns a morning read into a delegated list, sent where the person will see it. It runs in Brief's chat and takes a few minutes.
Before you start
Two things make this work better. Connect the chat tool your team lives in, under Settings → Connections, so the list can be delivered rather than copied by hand. And turn on Voice Print under Settings → Briefing & Actions, so what Brief writes sounds like you.
Step 1: Triage the brief against your priorities
Open the chat with ⌘⌥B and tell Brief what you are optimizing for. The brief ranks by what changed. Only you know what matters this quarter.
Here are my priorities for this quarter: [list them]. Go through my brief for today and tell me what actually needs me, what can wait, and what should belong to someone else.
The third bucket is the one to pay attention to. Work you have been carrying that has no reason to be yours tends to be invisible until something sorts for it.
Step 2: Ask for one person's list
Pick one person. A list built for a named person with a real role beats a general delegation plan.
Build a focus list for [name]. They joined three weeks ago and own operations. Ask me any questions you need before you write it.
That last sentence is the part people skip, and it is what makes the output usable. Brief will come back with questions instead of guessing: who owns the design once you approve the concept, whether that reimbursement is still outstanding, which folder they need access to.
Step 3: Answer the questions in one message
Answer all of them in a single reply, numbered to match. It takes a couple of minutes and it is the highest-leverage typing you will do that morning, because every answer is a thing your team would otherwise have interrupted you to ask.
Brief then produces the list with those answers built in.
Step 4: Send it where they read it
Send that list to [name] as a Slack DM.
Sending to another person waits for your approval by default. The action appears on the Actions page marked "Awaiting approval" with the message in front of you, and it goes out when you press Approve. See What Brief can do on your behalf for how to change that once you trust it.
If you would rather send from your own inbox, ask for a draft instead and it lands in your drafts folder.
Step 5: Make it recurring
If the sequence works once, it works every week, and you should not have to retype it.
Every Monday at 7am, do this: [paste the sequence]. Send me the result before you send anything to anyone.
That creates an agent, which is Brief running a job you defined on a schedule you set. Agents live on the Agents page, and you can change one by asking. See Video: Building agents.
Related skills
Brief has named skills that cover parts of this, and invoking one by name is faster than describing it:
- Triage Inbox sorts recent unhandled email into respond-now, FYI, and delegate.
- Prioritize forces a ranking on competing demands so you can decide what to drop, defer, or delegate.
- Prep for 1:1 pulls open commitments on both sides before you meet with someone.
The full list is in Skills and example prompts.
Doing this from ChatGPT or Claude
The same sequence runs in whichever assistant you already work in, with Brief connected as an MCP server. Your assistant pulls the briefing, the commitments, and the contacts from Brief, and reasons over them alongside whatever else you have given it.
One thing worth setting up first if you go this route: a project in your assistant that carries your role, your company, your market, and your current priorities as standing instructions. Then Step 1 stops being something you type every morning, because the assistant already knows what you are optimizing for and only needs the day's context from Brief. See Connecting to the MCP server, and Video: Connect AI assistants with MCP for it running.
One courtesy
When a message goes out under your name, say what sent it. A line at the top naming the assistant costs nothing and saves the person on the other end from reading your tone for clues. People care less about whether you used AI than about not having to guess.