PlainAgent / AI automation for beginners
Guide · AutomationAI automation for beginners
AI automation means adding an AI step to a workflow that runs on its own: when an email arrives, a model reads it and decides what happens next. It's the most practical place to start with AI at work because the steps are fixed and easy to check.
Automation, AI automation and agents
| Type | Who decides the steps | Example |
|---|---|---|
| Regular automation | You, with fixed rules | New form entry → add row to sheet |
| AI automation | You fix the steps; AI handles one fuzzy step | New email → AI labels it → route by label |
| AI agent | The AI picks which tools to use and in what order | "Research this lead" → searches, reads, writes a brief |
Start with AI automation. It's easier to test, and most "agents" in practice are AI automations with one or two decisions.
How to pick your first workflow
- It happens often (several times a week).
- It has a text step a person currently reads and judges.
- Mistakes are cheap or caught by review.
- You can check the result quickly.
Good first picks: labelling incoming email, extracting fields from invoices, turning meeting transcripts into tasks, summarising form submissions.
The tools
Zapier and Make are hosted and quick to start. n8n can be self-hosted and is popular for AI workflows. All three have AI steps and connect to common apps. Check current plans and limits on their sites.
A simple build pattern
- Trigger (new email, form, file, schedule).
- Get the data you need (lookups).
- AI step with a clear prompt and structured output (JSON).
- Route on the output (if/else).
- Action, starting with drafts or a review step.
- Log every run to a sheet so you can check it.
For importable workflows, see RecipeAgents n8n templates. For the decisions behind each one, see our agent templates.