Automate Your Life with AI lessons
What AI automation means
An AI sorts an inbox, drafts a report and fills in a spreadsheet, on a gentle loop.
AI automation means handing repetitive work to an AI tool. It can sort your email, draft documents and fill in spreadsheets.
You do not need to write code. You tell the tool what you want in plain words, then check what it gives back.
This course shows you how to pick the right tool, write clear instructions and turn everyday tasks into repeatable workflows.
Three ways AI tools work
AI tools work with you in three main ways. Popular names like ChatGPT, Claude, Gemini and Copilot now work in more than one of them.
A chat assistant replies to what you type in a chat window. A built-in assistant works inside an app you already use, like your email or a spreadsheet.
An agent works through several steps for you, such as searching, opening files and drafting. It reports back when it is done.
The three kinds of AI tool, each with an example request and what happens next.
Practice: Pick the kind of tool
Sort tasks by the kind of AI tool that suits them best.
Write it like a brief
An AI tool only knows what you tell it. A vague request gets a vague answer.
A clear instruction has four parts: the task, the context, the audience and the format. Think of it as briefing a new colleague.
An instruction gains its four parts one at a time, and the reply gets better with each.
Practice: Label the parts
Match each piece of an instruction to the part it plays.
Check it, then steer it
The first reply is a draft, not the finished job. Read it and spot what is missing or wrong.
Then send a follow-up that names the fix, like "shorter" or "as a table". The AI remembers the conversation, so you do not need to start again.
A chat where two follow-ups turn a vague summary into a useful one.
Practice: Pick the follow-up
Pick the follow-up that fixes each problem with an AI reply.
Save it as a template
A workflow is a task you do the same way every time. Once an instruction works, save it and reuse it.
Most assistants let you save instructions, for example in a project, a custom assistant or a saved prompt. Each time, you only add the new part.
One saved instruction drafts a reply to three different customer emails.
Practice: Build a reusable instruction
Fill the gaps in a saved instruction with the specific pieces, not the vague ones.
Email: let AI sort the inbox
Email is a good first workflow because it repeats every day. AI can label new messages, summarise long threads and draft replies.
Keep yourself as the last step. Read every draft before it goes out, because you are responsible for what you send.
An inbox gets labelled, sorted, summarised and given draft replies.
Practice: Triage the inbox
Label each email as urgent, needs a reply, or for your information.
Documents: summaries and action items
Long documents are where AI saves the most reading time. You can upload a file and ask for a summary, the key figures or a list of actions.
Ask it to quote the line each answer came from. Then you can check the parts that matter in seconds.
Meeting notes become a list of actions, and a quote shows where one came from.
Practice: Ask for the right thing
Pick the request that gets the job done for each document.
Spreadsheets in plain words
You can describe what you want from a spreadsheet in plain words. The AI writes the formula, cleans up messy data or suggests a chart.
Some spreadsheet apps also have an AI function you type into a cell, like a formula. Whatever the tool, check a few results by hand before you trust the whole column.
A plain request becomes a formula that fills a Total column, then highlights the big ones.
Practice: Check the AI's formula
Work out what a formula should show, to check the AI got it right.
Presentations: outline first
Do not ask for a whole deck in one go. First ask for an outline, a list of slide titles, and fix that.
Then ask the AI to draft each slide from your outline. Add your real figures and pictures last.
An outline is edited, then each slide is drafted from it.
Practice: Order the workflow
Put the steps of an AI workflow in order.
Research you can check
AI tools can search the web and pull many sources into one answer. Some have a research mode that digs deeper and lists its sources.
AI can still get facts wrong or misread a source. Open the sources behind the claims that matter and check they really say it.
Each claim in an AI answer is checked against its source.
Practice: Spot the claim to check
Tap the sentence in an AI answer that most needs checking.
Keep private data private
Check which AI tools your workplace allows before you paste in work. Approved work accounts often have stronger privacy rules than personal ones.
Leave out what the AI does not need, like names, contact details and account numbers. Swap them for placeholders and add the real details back yourself.
Personal details in a request are spotted and swapped for placeholders.
Practice: Safe to paste?
Sort each item by whether it is safe to paste into an AI tool.
Let an agent run the routine
An agent can carry out a whole workflow, like opening a sheet, filling in a report and drafting an email. Some assistants can also run a saved task on a schedule, such as every Monday morning.
Give an agent clear limits. Let it draft and prepare, and keep the final send or payment for yourself.
An agent prepares the weekly report and email, then stops for you to check.
What to automate first
Start with tasks you do often that follow the same pattern. These save the most time for the least risk.
Tasks that need real judgement, like hiring or a difficult conversation, should stay with people. AI can help you prepare, but you make the call.
Pick one task this week, write the instruction and save it. Improve it a little each time you use it.
Practice: Choose what to automate
Place each task by how often it comes up and how much judgement it needs.