Most AI assistants wait for you to ask them something. But what if your AI could work while you sleep—checking your email, summarizing news, organizing files, or preparing reports on a schedule?
This tutorial shows you how to set up automated AI tasks on Zo Computer. Once configured, these tasks run on their own, delivering results via email or SMS.
What you can automate
AI task automation works best for recurring work that follows a pattern:
Morning briefings — Get a daily summary of your inbox, calendar, or news topics
Research monitoring — Track competitors, industry news, or topics you care about
File organization — Automatically sort downloads, rename photos, or clean up folders
Report generation — Weekly summaries of metrics, activity logs, or project status
Content curation — Collect and summarize articles, papers, or social media posts
The key is identifying tasks you do repeatedly that could run without your input.
Setting up your first automated task
On Zo Computer, automated tasks are called Agents. Each agent has an instruction (what to do) and a schedule (when to do it).
Step 1: Navigate to Agents
Go to Agents in your Zo Computer dashboard. You'll see any existing scheduled tasks and a button to create new ones.
Step 2: Write your instruction
Be specific about what you want. Good instructions include:
What to do: The actual task
Where to get information: Files, websites, connected apps
What to deliver: The format and level of detail you want
Example — Daily news summary:
Check the top stories on Hacker News and TechCrunch. Summarize the 5 most interesting stories about AI in 2-3 sentences each. Focus on practical applications, not hype.
Example — Weekly email digest:
Review my Gmail inbox from the past week. Identify emails that need responses but haven't been answered. List them with the sender, subject, and a one-line summary of what they're asking.
Step 3: Set the schedule
Choose when and how often the task runs:
Daily at 7am — Good for morning briefings
Every Monday at 9am — Weekly summaries
Every hour — High-frequency monitoring (use sparingly)
Step 4: Choose delivery method
Decide how you want results delivered:
Email — Best for longer content, reports, or anything you'll reference later
SMS — Best for short alerts or summaries you need immediately
Real examples
Morning briefing
Instruction:
Every morning, check my Google Calendar for today's meetings. For each meeting, find the attendees' LinkedIn profiles and summarize their background in one sentence. Also check if I have any unread emails from those people.
Schedule: Daily at 6:30am
Delivery: Email
Competitor monitoring
Instruction:
Search for news about [competitor name] from the past 24 hours. Summarize any product announcements, funding news, or significant press coverage. If nothing notable happened, just say "No significant updates."
Schedule: Daily at 8am
Delivery: SMS
Weekly file cleanup
Instruction:
Look through my Downloads folder. Move any PDFs to Documents/PDFs, images to Pictures, and delete any .dmg or .exe files older than 7 days. Tell me what you moved and deleted.
Schedule: Every Sunday at 10pm
Delivery: Email
Tips for reliable automation
Start simple. Get one task working before adding complexity. A daily weather summary is a better first agent than a complex multi-step workflow.
Be explicit about output format. "Give me a bullet list" or "Keep it under 200 words" helps ensure consistent results.
Test manually first. Before scheduling, run the instruction yourself in chat. If it works interactively, it'll work as an agent.
Set realistic schedules. Tasks that query external services or process lots of data take time. Don't schedule them every 5 minutes.
Check your agents periodically. Review the Agents page to see run history and catch any failures.
What's happening under the hood
When an agent runs, Zo executes your instruction with full access to:
Your connected apps (Gmail, Calendar, Drive, etc.)
Web browsing and search
Your files and folders
Any tools or scripts on your Zo
The agent runs in its own session, completes the task, and delivers results through your chosen method. You can see logs and history in the Agents dashboard.
Next steps
Once you're comfortable with basic agents, try:
Chaining agents — One agent's output becomes another's input
Conditional logic — "Only notify me if X happens"
File-based workflows — Agents that read and write files for persistent state
The goal is to make your AI genuinely useful without constant supervision. Start with one task you do every day, automate it, and build from there.