Lesson 4 of 5 · about 5 min

Simple automations that connect apps

When you finish, you can plan an automation that moves information between apps, uses an AI step, and includes safety checks.

Stuck? Ask the AI tutor

The trigger, the steps and the result

An automation is a chain: something happens, then steps run, then a result lands somewhere. This is often written as trigger, action, destination. A new form answer arrives, an AI step drafts a summary, and the summary is added to a spreadsheet row or sent to your email.

Several services let you build these chains by clicking and filling in boxes, and many AI products offer some built-in automation. The names and limits change, so check current help pages. What matters is that you can describe the chain clearly before you build it.

Draw it first

Before opening any tool, write the chain on paper in plain sentences. “When someone submits the request form, the AI writes a two-line summary and a suggested priority. The summary and priority are added to my tracking sheet. I get an email to review.” This takes five minutes and prevents most confusion later.

Then mark the risk points. What if the form is empty? What if the AI answers oddly? What if the sheet is full or the link breaks? For each risk, decide: stop and alert me, or continue with a note. Good automation expects trouble.

Permissions and costs

Each connection asks for permission to read or write in another app. Give only what is needed. If the automation only needs to add rows to one sheet, do not give it access to your whole drive. Remove connections you no longer use.

Some automation and AI services charge by usage, and a loop that runs too often can cost more than you expect. Start small, watch the usage, and set limits if the service offers them. Never put passwords or secret keys into a prompt or share them in a message.

Worked example: weekly feedback digest

A small bookshop in Cape Town collects customer comments by form. Each Friday, an automation gathers the week’s comments, asks the AI to group them into themes and list any urgent complaints, and emails the owner a short digest. Nothing is sent to customers.

The owner reads the digest over coffee and opens the original comments for anything urgent. She checks the AI’s themes against a few real comments the first few weeks. The automation saves her an hour, and the safe design means a wrong theme costs little.

Practice now

  1. Write a trigger, steps and destination for one small task you do weekly. Mark where a person should check the result and what should happen if input is missing.
  2. List which apps the automation would need access to, and write down the narrowest permission each one needs.

Use any AI assistant you have. Do not paste private, client or patient information into it.

Check yourself

1. What is a good first step in planning an automation?

2. How should you handle permissions for connected apps?

3. Why is a digest sent only to the owner a safe first automation?

Ask the AI tutor

Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.

AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.