Day 17 of 28 · OpenClaw Challenge
Automating Recurring Tasks
⏱ 7 min
📊 Intermediate
Until now, your agent has been reactive — you ask, it does. That's already powerful.
But the real game-changer is when your agent works proactively — doing things on a schedule without you ever asking.
Reactive vs proactive
Reactive: You say "check my email." The agent checks your email.
Proactive: Every morning at 7am, the agent checks your email, sorts it, drafts replies, and sends you a summary. You didn't ask. It just does it.
This is the shift from "assistant you talk to" to "assistant that works for you in the background."
Knowledge Check
What's the difference between reactive and proactive agent behavior?
A
Reactive waits for your command — proactive runs on its own based on triggers you've set
B
Proactive costs more money
C
Reactive is faster than proactive
D
There's no real difference
Reactive means the agent waits for you. Proactive means the agent acts on its own, based on schedules or events you've configured. Your morning briefing from Day 14? That's proactive.
Three types of triggers
You can make your agent act based on three kinds of triggers:
Time-based — "Every morning at 8am," "Every Friday at 5pm," "Every hour during business hours"
Event-based — "When a new email arrives from a client," "When a Slack message mentions my name," "When a PR is merged"
Condition-based — "When the server error rate exceeds 1%," "When my inbox has more than 50 unread emails," "When a deadline is less than 24 hours away"
Knowledge Check
Which type of trigger would you use for "alert me when our website response time exceeds 2 seconds"?
B
Condition-based — it fires when a metric crosses a threshold
Condition-based triggers monitor a value and fire when it crosses a threshold. "Response time > 2 seconds" is a condition. The agent continuously monitors and acts only when the condition is met.
Real automations people use
These aren't hypothetical. These are recurring tasks that OpenClaw users set up every day:
Set up a daily standup summary. Every day at 9:30am, read everyone's standup updates in #standups and post a digest.
Daily standup automation configured ✓
Every day at 9:30am, I'll:
1. Read all updates posted in #standups since yesterday
2. Group them by team member
3. Highlight blockers and dependencies
4. Post a digest in #engineering-leads
First one runs tomorrow at 9:30am. Want to preview what it'll look like?
Here's what today's would look like:
📊 **Standup Digest — Feb 19**
**Sarah**: Finished auth refactor. Starting API docs today.
**Jake**: Bug fix deployed. Picking up dashboard ticket.
**Marcus**: Blocked on design assets from Priya.
**Priya**: Design review with client today. Will send assets to Marcus by EOD.
🚩 **Blocker**: Marcus → waiting on Priya (ETA: end of today)
Looks good?
↻ Replay conversation
Knowledge Check
What three things does the standup automation do beyond just reading updates?
A
It schedules 1:1 meetings for blocked team members
B
It sends individual reminders to each person
C
It groups by person, highlights blockers, and posts a digest to a leadership channel
D
Nothing — it just copies the messages
The automation doesn't just copy — it synthesizes. It groups by person (organization), identifies blockers (analysis), and posts to the right channel (distribution). That's the difference between automation and intelligence.
Stacking automations
Once you're comfortable with one automation, start stacking them:
Morning: Inbox triage + standup digest + briefing
During work: Monitor Slack for urgent mentions + watch for new client emails
End of day: Summarize what you accomplished + prep tomorrow's agenda
Weekly: Generate a progress report from the week's activity
Each automation is simple on its own. Together, they create an always-working assistant that handles the repetitive parts of your job.
How many automations do I have running?
You have 4 active automations:
1. ☀️ **Morning briefing** — daily at 7am
2. 📊 **Standup digest** — daily at 9:30am
3. 📧 **Client email alerts** — event-based (real-time)
4. 📋 **Weekly summary** — Fridays at 4pm
All running smoothly. Last failure: none. Want to add or modify any?
↻ Replay conversation
Final Check
What's the risk of setting up too many automations at once?
A
Each automation costs a separate monthly fee
B
You might get overwhelmed by alerts and lose track of what the agent is doing — start simple and add gradually
C
There's no risk — more is always better
Start with one or two automations and build up. If you set up 15 automations on day one, you'll get flooded with notifications and lose trust in the system. Add gradually, verify each one works, then stack.
🔄
Day 17 Complete
"Your agent now works even when you're not watching. Set the triggers, let it run."
Tomorrow — Day 18
Memory Management
Your agent is doing more than ever. But is it remembering everything it should? Let's find out.