Without Scheduled Tasks, your AI employees are reactive. They sit idle until you open a conversation and ask them to do something. That's better than a generic chatbot, but it's still assistant behavior.
With Scheduled Tasks, your AI employees become proactive. They wake up at 8 AM and review the sales pipeline. They produce a weekly marketing report every Monday. They check data quality every Friday afternoon. They do their job without being asked, because that's what employees do.
What Are Scheduled Tasks?
Scheduled Tasks is a feature in Claude Desktop that lets you set up prompts to run automatically on a schedule. You define:
When the scheduled time arrives and your computer is on, Claude Desktop automatically runs the prompt, processes it with the full Project context, and saves the output. You check it when you're ready.
Why This Is the Game Changer
Consider the difference:
Without Scheduled Tasks (reactive)
- Monday morning. You remember you need a content plan.
- You open Claude. You navigate to your Marketing Strategist project.
- You type: "Create a content plan for this week."
- You wait for the output. You review it. You give feedback.
- Total time: 20 minutes of your active involvement.
With Scheduled Tasks (proactive)
- Monday morning. You open your laptop.
- A notification shows your Marketing Strategist has already produced the weekly content plan.
- You review it over coffee. Five minutes. Done.
- Total time: 5 minutes of your active involvement.
The time savings compound. Multiple AI employees, each running 1-3 scheduled tasks per week. That's 10-20 deliverables produced without you initiating anything. Your role shifts from "person who prompts AI" to "manager who reviews AI output." That's the employee relationship.
How to Set Up Scheduled Tasks
Step 1: Open Claude Desktop Settings
- Open the Claude Desktop application
- Click the menu icon or navigate to Settings
- Look for Scheduled Tasks or Automations in the settings menu
Step 2: Create a New Scheduled Task
- Click Add Scheduled Task (or equivalent button)
- Give it a descriptive name: "Marketing Strategist - Weekly Content Plan"
- Select the Project this task should run in (your Marketing Strategist project)
- Write the prompt that will execute on schedule
Step 3: Write the Prompt
The prompt for a Scheduled Task should be self-contained and specific. Don't write "help me with marketing." Write exactly what the employee should produce:
It's Monday morning. Perform your Weekly Content Planning workflow:
1. Review the content calendar for upcoming deadlines this week
2. Identify any gaps or overdue items
3. Draft the content plan for this week including:
- Article topics and titles (3-5 pieces)
- Target keywords for each piece
- Priority levels (high/medium/low)
- Suggested publication dates
- Brief outline (3-5 bullet points per piece)
4. Flag any items that need my input or approval
5. Note any industry trends worth covering this week
Format the output as the Weekly Content Plan using the standard template.
Step 4: Set the Schedule
Choose the frequency and timing:
- Daily tasks: Set the time (e.g., 8:00 AM)
- Weekly tasks: Set the day and time (e.g., Monday at 9:00 AM)
- Custom schedules: Specific days of the week, specific times
Step 5: Save and Verify
Save the Scheduled Task. Verify it runs correctly by either waiting for the next scheduled time, or manually triggering it to test.
Suggested Schedules by Role
These are starting points. Adjust based on your actual workflows.
Marketing Strategist
| Task | Schedule | Description |
|---|---|---|
| Weekly Content Plan | Monday, 9:00 AM | Plan the week's content with topics, keywords, and briefs |
| Performance Review | Friday, 3:00 PM | Analyze the week's marketing metrics and identify trends |
| Competitor Scan | Wednesday, 10:00 AM | Research competitor activity and market movements |
| Monthly Strategy Review | 1st of month, 9:00 AM | Comprehensive monthly marketing performance analysis |
Sales Development Rep
| Task | Schedule | Description |
|---|---|---|
| Daily Lead Review | Every day, 8:00 AM | Review new leads, prioritize outreach, update pipeline |
| Outreach Sequence Check | Every day, 2:00 PM | Review follow-up sequences, draft next-step emails |
| Weekly Pipeline Report | Friday, 4:00 PM | Summarize pipeline status, win/loss analysis, forecast |
| ICP Refinement | Monthly, first Monday | Review closed deals and refine ideal customer profile |
Customer Support Agent
| Task | Schedule | Description |
|---|---|---|
| Ticket Queue Review | Every day, 8:00 AM | Triage open tickets, identify urgents, draft responses |
| Knowledge Base Audit | Friday, 2:00 PM | Identify gaps based on the week's tickets |
| Weekly Support Report | Friday, 4:00 PM | Ticket volume, resolution times, common issues, trends |
| FAQ Update | Monthly, last Friday | Update FAQ based on the month's most common questions |
Content Writer
| Task | Schedule | Description |
|---|---|---|
| Daily Writing Session | Every day, 9:00 AM | Draft or revise the current priority piece |
| Research Roundup | Monday, 8:00 AM | Compile research and inspiration for the week's writing |
| Editorial Review | Thursday, 3:00 PM | Review all drafted content for consistency and quality |
| Style Guide Check | Monthly, first Monday | Review recent output against brand voice and style |
Data Analyst
| Task | Schedule | Description |
|---|---|---|
| Daily Dashboard Check | Every day, 8:00 AM | Review key metrics, flag anomalies |
| Weekly Report | Friday, 4:00 PM | Comprehensive weekly data analysis with trends |
| Data Quality Audit | Wednesday, 10:00 AM | Check data sources for completeness and accuracy |
| Monthly Deep Dive | Last Friday, 9:00 AM | Deep analysis on a specific topic (rotating focus) |
Executive Assistant
| Task | Schedule | Description |
|---|---|---|
| Daily Briefing | Every day, 7:30 AM | Today's schedule, priorities, preparation notes, reminders |
| Weekly Agenda Prep | Sunday, 8:00 PM | Next week's calendar review, conflict identification |
| Action Item Tracker | Friday, 5:00 PM | Compile outstanding action items from the week |
| Monthly Calendar Review | Last Friday, 3:00 PM | Next month outlook, scheduling optimization |
Operations Manager
| Task | Schedule | Description |
|---|---|---|
| Daily Operations Check | Every day, 8:30 AM | Cross-role status check, bottleneck identification |
| Process Efficiency Report | Friday, 3:00 PM | Weekly operations summary, process improvements |
| Cross-Role Coordination | Wednesday, 10:00 AM | Review handoffs between roles, identify gaps |
| Monthly Operations Review | 1st of month, 10:00 AM | Comprehensive operations health check |
Power Patterns
Cross-Agent Workflows
Scheduled Tasks run as separate conversations, so one agent can't directly see another's output. The solution: agents write their output to local files, and downstream agents read those files.
For example, your Marketing Strategist runs at 9am and the prompt includes "Save the content plan to ~/ai-outputs/content-plan.md." Your Content Writer runs at 10:30am and the prompt begins with "Read ~/ai-outputs/content-plan.md and use it as the basis for today's blog draft."
This gives you a real team workflow where each agent builds on the last one's work. The files on your computer act as the handoff.
Recommended conventions for cross-agent file handoffs:
- Use a dedicated output folder like
~/ai-outputs/so all agent work lives in one place - Name files by role and deliverable, e.g.,
marketing-content-plan.md - Include a timestamp in the file so downstream agents know if the data is fresh
- Schedule agents 15-30 minutes apart when one depends on another's output
Using Memory for Continuity Between Runs
Scheduled Tasks don't automatically have memory of previous runs. To create continuity:
This is just one chapter.
Before you can set up Scheduled Tasks, you need an AI employee to schedule. The Playbook covers everything from scratch: how to design a role identity, build a memory system, create workflows, set guardrails, and deploy your first AI employee step by step. This chapter builds on that foundation.