It’s 6:47 AM. Your alarm hasn’t gone off yet, but you’re already
half-awake, reaching for your phone the way everyone does — muscle
memory, zero thought. You open Telegram. There’s a message waiting from
your AI, sent twelve minutes ago, while you were still dreaming.
“Good morning. It’s 7°C and overcast — rain expected after 2 PM,
so grab a jacket if you’re heading out later. You’ve got three meetings
today: the product sync at 10, a one-on-one with Sarah at noon, and the
investor prep call at 4. That last one got moved from 3 — check your
calendar. Two emails worth reading: a contract revision from legal (they
changed the liability clause again) and a shipping notification for the
monitor arm you ordered. Sterling is down slightly against the dollar.
And the Guardian is running a piece on the AI regulation bill that’s
relevant to your pitch deck. Have a good one.”
You didn’t ask for any of this. You didn’t type a prompt. You didn’t
open six apps and piece together your day like a jigsaw puzzle. Your AI
did it for you, unprompted, on schedule, every single morning.
This isn’t a concept demo. This isn’t a “coming soon” feature from a
startup’s landing page. This is what happens when you set up an
ai agent for personal use that actually runs in the
background — not a chatbot you poke when you remember it exists, but a
persistent assistant that works while you sleep.
Here’s how to build one.
What a Morning
Briefing Actually Looks Like
Before diving into configuration, it’s worth understanding what makes
a good ai daily briefing different from a notification
dump. Your phone already screams at you. Email badges, calendar alerts,
news push notifications, weather widgets — they’re all competing for the
same slice of attention, and none of them talk to each other.
A proper morning briefing is synthesised. It’s one message, in one
place, written in natural language by an AI that knows what matters to
you. The weather isn’t just a temperature — it’s a recommendation. The
calendar isn’t just a list — it’s context about what changed overnight.
The emails aren’t all of them — just the ones that actually need your
brain.
The difference between “information” and “a briefing” is editorial
judgement. And that’s exactly what a personal ai
assistant brings to the table.
The Building Blocks: What
to Include
Every good briefing pulls from a handful of sources. You don’t need
all of them on day one — start with two or three and expand as you get
comfortable.
Weather and location context — Temperature,
precipitation, and a practical note (“bring an umbrella” or “good day
for a walk”). If you travel frequently, your AI can pull weather for
wherever you actually are.
Calendar overview — Today’s events with start times,
plus anything that changed since yesterday. Moved meetings, new invites,
cancellations. The delta matters more than the static list.
Email triage — Not every email. The important ones.
A good prompt tells your AI to surface emails from specific senders,
emails with attachments, emails that look like they need a reply, and to
ignore newsletters and marketing. Think of it as an executive summary of
your inbox.
News that’s relevant to you — Not a generic news
feed. Curated headlines based on your interests, your industry, or
topics you’ve told your AI to track. “AI regulation news” is more useful
than “top stories.”
Portfolio or financial snapshot — If you track
investments or currencies, a quick line on how things moved overnight.
Nothing detailed — just enough to know if something needs attention.
Personal reminders — Anything your AI knows about: a
package arriving, a bill due date, someone’s birthday. The small stuff
that falls through the cracks.
Setting It Up with OpenClaw
OpenClaw is the engine that makes this work. It gives your AI a
persistent home — a place to run, remember things, and reach you on your
preferred channel. If you’ve already got an agent running (maybe you’re
using clawdbot on Telegram, or you’ve set up
moltbot for your Discord server), you’re halfway
there.
The core mechanism is a cron job — a scheduled task
that fires at a specific time and tells your AI to compose and send your
briefing. If you’re not familiar with how OpenClaw handles scheduling,
the cron jobs and
automations guide covers everything in detail. But here’s the short
version.
Choosing Your Channel
Your briefing needs to land somewhere you’ll actually see it.
OpenClaw supports several messaging platforms, and which one you pick
matters more than you’d think.
Telegram is the most popular choice for personal
briefings. It’s fast, notifications are reliable, messages support rich
formatting, and you can set up a dedicated bot that feels like texting a
very competent friend. Most people running an openclaw morning
briefing use Telegram as their primary channel.
WhatsApp works well if that’s where you already
live. The integration means your briefing shows up alongside your normal
messages, which can be either a feature or a bug depending on how you
feel about mixing AI messages with group chats.
Discord is ideal if you’re already running a
personal server or want to keep AI interactions in a dedicated space.
It’s also the natural fit if you’re using moltbot or
clawdbot in a community server and want to extend that
to personal use.
Pick the one you check first in the morning. That’s your channel.
The Cron Configuration
Here’s a real cron setup that delivers a weekday morning briefing at
7 AM:
name: morning-briefing
schedule: "0 7 * * 1-5"
channel: telegram
model: claude-sonnet-4-20250514
prompt: |
You are my personal morning briefing assistant. Compile and deliver
a concise morning briefing. Use a warm but efficient tone — like
a thoughtful chief of staff, not a news anchor.
Check and include:
- Weather for London today (temperature, conditions, practical advice)
- My calendar events for today, noting any changes from yesterday
- Important unread emails (skip newsletters and marketing)
- Relevant news headlines for: AI industry, UK tech policy, startup funding
- GBP/USD and GBP/EUR exchange rates
- Any reminders or notes I've left for today
Format as a single flowing message. No bullet points in the output,
no greetings like "here's your briefing" — just start talking as if
we're already mid-conversation. Keep it under 300 words.That 1-5 in the schedule means Monday through Friday.
The model choice matters — you want something smart enough to synthesise
but not so expensive that your daily briefing costs more than your
morning coffee. Running your entire ai agent for personal
use setup for under
five dollars a month is entirely realistic, briefings included.
The Prompt Is Everything
Notice how specific that prompt is. It doesn’t say “give me a
briefing.” It says what to check, how to write, and
what to skip. The difference between a useful briefing and a
generic one lives entirely in the prompt.
A few principles worth following:
Tell it what to ignore. “Skip newsletters” saves you
from a daily summary of marketing emails you were never going to read
anyway. Exclusion rules are as important as inclusion rules.
Set a tone. “Like a thoughtful chief of staff” gives
the AI a persona to work with. Without tonal guidance, you’ll get
something that reads like a Wikipedia summary — accurate but
lifeless.
Cap the length. Three hundred words is about a
ninety-second read. Long enough to be useful, short enough that you’ll
actually read it every day. A briefing you skip is worse than no
briefing at all.
Reference specific interests. “AI industry, UK tech
policy” is better than “news.” The more specific you are, the more
relevant the output.
Going Further:
Weekday vs Weekend Briefings
Here’s where things get interesting. Your Monday morning and your
Saturday morning are fundamentally different. One needs meeting prep and
email triage. The other might benefit from a weather-optimised activity
suggestion and a reminder that you promised to call your parents.
Set up a second cron job for weekends:
name: weekend-briefing
schedule: "0 9 * * 0,6"
channel: telegram
model: claude-sonnet-4-20250514
prompt: |
Weekend morning briefing. Keep it relaxed and short.
Include:
- Weather for today and tomorrow (activity-oriented: "good day for
a run" or "stay in and read" kind of vibe)
- Any calendar events (rare on weekends, but check)
- One interesting long-read or article from this week I might enjoy
(topics: AI, space, economics, history)
- Any personal reminders for the weekend
Tone: friendly, unhurried. Like a flatmate who read the paper
before you woke up. Under 200 words.Notice the differences. Later send time — 9 AM instead of 7.
Different tone. Different content priorities. Shorter length. The
schedule (0,6) targets Saturday and Sunday
specifically.
This kind of contextual awareness is what separates a tool from an
assistant. Anyone can set up a weather notification. An ai agent
for personal use understands that the same person needs
different things on different days.
Using Heartbeats for
Smarter Timing
Cron jobs are great for fixed schedules, but OpenClaw also supports
heartbeats — periodic check-ins where your AI can decide whether
something is worth sending. Think of heartbeats as giving your assistant
discretion.
Instead of blasting a briefing at 7 AM regardless, a heartbeat-based
approach lets your AI check: is there actually anything worth
reporting this morning? If your calendar is empty, no important
emails came in overnight, and the weather is unremarkable, maybe today’s
briefing is just: “Nothing urgent. Enjoy the quiet morning.”
You can combine both approaches — a cron job for the reliable daily
briefing, and heartbeats throughout the day for things that come up. An
urgent email at 2 PM. A meeting that just got cancelled. A stock that
dropped more than your threshold. The automations guide
walks through heartbeat configuration in detail.
Making It Yours
The examples above are starting points. The real magic happens when
you tune the system over a few weeks. You’ll notice patterns — maybe you
never read the financial section, so you drop it. Maybe you want a “this
day in history” fact because it gives you something interesting to
mention in meetings. Maybe you want your AI to flag when you have a free
afternoon so you can book a gym slot.
Some people running clawdbot on Telegram have set up
briefings that include package tracking, public transport disruptions on
their commute route, and even a daily vocabulary word in a language
they’re learning. Others keep it minimal — weather, calendar, done.
The point isn’t to build the most comprehensive briefing possible.
It’s to build the one you’ll actually read every morning. Start
lean. Add things when you feel something’s missing. Remove things when
you start skipping over them.
The Bigger Picture
A morning briefing is usually the first automation people set up when
they start using an ai agent for personal use, and
there’s a reason for that. It’s simple, it’s immediately useful, and it
changes how your morning feels.
But it’s also a gateway. Once you experience what it’s like to have
an AI that does things without being asked, you start seeing
opportunities everywhere. Evening summaries. Weekly reviews. Automated
email drafts. Research that happens overnight so the results are waiting
when you sit down to work.
The morning briefing is the proof of concept. It’s the thing that
makes you think: okay, this actually works. What else can it
do?
And the answer, increasingly, is: quite a lot.
Whether you’re running moltbot in a Discord
community or a private clawdbot instance on Telegram,
the underlying principle is the same. Give your AI a schedule, a clear
set of instructions, and a channel to reach you. Then let it do its job
while you do yours.
Your alarm goes off tomorrow at 6:45. By the time you pick up your
phone, your briefing will already be there. You didn’t build a chatbot.
You built a butler.
And honestly? You’ll wonder how you ever started your day without
one.