How to Build an AI Morning Briefing That Replaces Six Tabs

An AI morning briefing is one scheduled message that arrives before you open your laptop, carrying today's calendar, overnight mail from people you named, and your top priorities from a file. In Zo Computer you build it by writing a prompt with a time, a word cap, and a delivery channel. Five minutes to read. Six tabs you skip.

Key takeaways

• An AI morning briefing in Zo is a Zo Computer automation: a plain-English prompt on a schedule, with the same files and connected apps Zo uses in chat.
• Connect Google Calendar and Gmail first. Those two carry almost everything a founder's brief needs.
• Zo cannot guess who your VIPs are. Name the three people and the one domain you can't miss, or the brief invents its own definition of urgent.
• Put a word cap in the prompt. Without one you get 600 words and quit reading by Thursday.
• Add a conditional clause so the brief stays silent on an empty Saturday. That line is the difference between an automation you keep for a year and one you mute in a week.
• Weather belongs in the brief only on travel or outdoor days. Zo reads public forecast pages with nothing connected.

What is an AI morning briefing?

An AI morning briefing is an AI daily digest that pulls your schedule, your inbox, and your stated priorities into one message at a fixed time. Most versions are a daily calendar and email summary with a priorities line on top. Summarizing is not the point. Elimination is. You stop opening the calendar tab, then Gmail, then the doc where your quarter goals live, then a weather app. The most-cited number on that loop comes from the McKinsey Global Institute report The Social Economy (July 2012): the average interaction worker spends about 28 percent of the workweek managing email. It is a 2012 figure and McKinsey has not republished a comparable benchmark, which is exactly why every article in this niche still quotes it.

In Zo, the brief is an automation: a prompt running on a schedule with your files, tools, and connected apps. Same Zo you talk to at 3pm.

What goes in the brief, and what stays out?

The failure mode of every first brief is stuffing. Everything goes in, the message runs long, and what you needed sits in paragraph four. Four sections, hard cap.

People skip the priorities line, and it is the one that turns a status report into something with an opinion. Keep a short file in your workspace: priorities.md, three bullets, updated Sunday. Point the prompt at it. Zo reads your files during the run, so it works with nothing connected.

Weather gets a note. Zo reads public web pages with no plugin and no API key, so a forecast costs nothing. But daily weather in a brief you read at a desk is filler, and filler is how a brief loses your trust. Gate it: "include the forecast only if I have travel or an outdoor commitment."

Which plugins do you connect first?

Two. Google Calendar and Gmail. Both sit in Zo's verified integration tier, so they behave more predictably than the 1,000-plus apps in the wider catalog.

Connecting takes about a minute. Ask Zo to read your calendar, a Connect card appears inline, you click it, you approve the OAuth window, the card flips to connected. Then the step that catches everyone: clicking Connect does not wake Zo back up. Send one more message. Say "go." Personal Gmail and work Gmail on one account? Name which you mean. A brief built on the wrong inbox is worse than no brief.

What does the starter prompt look like?

This prompt is a starting point, not a fixed script. Every part of it is meant to be edited: the time, the timezone, the names, the file, the word cap, the sections themselves. If you do not use one of the sources below, delete that line and write the one you do use. Zo reads the prompt as plain instructions, so changing the words changes what it does.

Every weekday at 6:45am Eastern, read my Google Calendar for today and my Gmail from the last 16 hours. Email me a brief with three sections:

1. TODAY: every meeting with its start time, and who it's with.
2. OVERNIGHT: any email from high level people. Sender and one-line gist. Skip newsletters,
receipts, and automated mail.
3. PRIORITIES: read priorities.md in my workspace and list the top three
verbatim.

Hard cap: 200 words. If a section is empty, write "nothing" and move on. Do not editorialize and do not add a section I didn't ask for.

Four things in there are doing real work. A 16-hour lookback catches the 9pm email without dragging in yesterday morning. Named VIPs replace Zo's guess with your judgment. The 200-word cap keeps it readable at a kitchen counter. And "write nothing and move on" stops Zo padding an empty section with an explanation of the emptiness.

Schedules go in as plain English and Zo converts them to a recurrence rule, confirming before it saves. Say "every weekday," not "every day," unless you want a Saturday brief. That mix-up is the most common scheduling mistake people make, and it bills you every weekend for nothing.

Which channel should the brief arrive on?

Zo delivers to email, SMS, Telegram, Discord, or Slack. By default an automation runs quietly and writes to chat history, so delivery is opt-in. Pick on how you read, not which app you like.

My call: SMS if the brief is genuinely 100 words, email if it is 200. Split delivery beats either. Route the routine brief to email, and text yourself only when something specific breaks. One run, two channels, one line of prompt.

How do you keep it quiet on empty days?

Add a conditional clause. The automation still runs, still reads the calendar and inbox, and stays silent unless the condition holds.

Every weekday at 6:45am Eastern, read my Google Calendar for today and my Gmail from the last 16 hours, plus priorities.md in my workspace.

Send the brief only if at least one of these is true: I have a meeting today, mail arrived from a high leverage contact, or something looks genuinely time-sensitive. If none of that is true, stay silent and log the run.

Format: TODAY (meetings with times), OVERNIGHT (sender plus one line),
PRIORITIES (top three from the file, verbatim). Include today's forecast for my location only if I have travel or an outdoor commitment on the calendar. Hard cap 200 words. Text me if the whole brief fits in 100 words, otherwise email it.

That is the upgraded version, and the one you should end up running. What does the quiet cost you? Nothing. Same schedule, same read, fewer interruptions. An alert you swipe away is worse than none, because you paid credits for it and it taught you to stop looking. Once it feels solid, extend it to meeting prep: same run, searching your inbox for the last thread with each external attendee and summarizing where you left off in two lines.

What goes wrong in week one?

Five things, in the order they usually bite.

Every run bills, condition met or not. Conditional delivery controls whether Zo messages you, not whether it runs. A five-minute schedule fires 288 times a day and bills for all 288 even if it texts you once a month. One run per weekday is all a brief needs. The pricing page covers plans and credits.

"Urgent" is your judgment, not Zo's. Leave the VIP list out and Zo picks its own definition of important, which buys you false positives all week. Prompt problem rather than model problem, fixed by naming people and domains instead of adjectives.

A lapsed token fails silently. The nastiest one. Your Gmail authorization expires, the schedule keeps firing on time, the app call fails, and the brief arrives looking normal with an empty OVERNIGHT section. You read "nothing" and believe it. If a section goes quiet three days running, reconnect before you assume your inbox got peaceful.

Fuzzy instructions produce fuzzy output. "Anything time-sensitive" leans on Zo's judgment, fine for triage and bad for anything involving money. Where a hard rule is writable, write it.

Zo drafts. It does not send. Extend the brief to suggested replies and the human gate stays. Nothing reaches a customer without you reading it first.

Every run is saved as a real conversation you can open, so an empty section is not something to theorize about. Tell Zo "open this morning's run of my brief and tell me why TODAY was blank," and it reads its own transcript.

How do you tune it across the first week?

Run it five days before changing anything structural, then sit with all five messages and make three edits.

1. Cut one section. There is always one you skipped every single morning, so delete it from the prompt and stop paying for it.
2. Fix the VIP list. Add whoever wrote and got missed, and cut whoever generated a false positive.
3. Move the time by 15 minutes. Most people set 7:00am and then actually read the thing at 6:40 or 7:20, so match the message to the morning you really have.

That is the loop. One brief, tuned twice, then left alone for months. Build this morning brief automation before anything else in your stack, because you feel it every day, and feeling it is what makes the next five obvious.

Frequently asked questions

What is an AI morning briefing?

An AI morning briefing, also called an AI daily briefing, is a scheduled message that consolidates your calendar, overnight email, and stated priorities into one short read at a set time. In Zo Computer it is an automation: a plain-English prompt running on a schedule with your connected apps and workspace files.

What should a morning brief include?

Today's calendar with times, overnight mail from senders you name explicitly, your top two or three priorities from a file you maintain, and weather only on travel or outdoor days. Cap it at 200 words. Leave out inbox counts, newsletters, and any list Zo generated itself.

How do I set up an automated daily briefing in Zo?

Connect Google Calendar and Gmail, then tell Zo in plain English when to run and what to send. Zo converts your sentence into a recurrence rule and confirms it before saving. You never write cron syntax or touch a raw rule.

Does an AI morning briefing cost credits every day?

Yes. Every scheduled run spends AI credits the same way a chat does, whether or not it notifies you. One run per weekday is cheap. A brief polling every 15 minutes is roughly 96 sessions a day, which is why interval length matters.

Can Zo include the weather without connecting an app?

Yes. Zo reads public web pages with nothing connected, so a forecast needs no plugin, no API key, no setup. Gate it in the prompt to travel days and outdoor commitments so it does not become filler.

Should the brief arrive by text or email?

Text it if it fits in a glance, and name the length you want in the prompt, because Zo aims short over SMS rather than enforcing a fixed cap. Email it if it carries meeting details you will scroll. You can also split delivery in one prompt: routine brief to email, urgent exception by SMS.