How To Learn AI: A 4-Week Plan Built On Your Real Work
The short version
- The fastest way to learn is to use AI on your own work, every weekday, for about 20 minutes.
- Spend four weeks on four jobs: writing, thinking, research, and building one small repeatable helper.
- Keep a simple log of what worked and what didn’t. That log becomes your personal prompt library.
- By the end of week four you should have at least five tasks you now do faster, with prompts saved for each.
- Skip theory until you’ve got a reason to care about it. Practice first, explanations second.
If you want to know how to learn AI as a busy adult, here’s the answer: pick real tasks from your own week and do them with AI, on purpose, for four weeks. You’ll learn faster than any course can teach you, because every lesson lands on work you already understand and care about.
Below is the four-week plan. Each week has one focus, a daily habit, and a finish line. You need one free chat tool and a notes file. That’s it. Most advice on how to learn AI starts with theory. This plan starts with your inbox.
Before Week One: Set Up In 15 Minutes
- Pick one chat assistant. ChatGPT, Claude, or Gemini all work on their free tiers. Don’t bounce between them yet.
- Make a note called AI Log. Each entry gets three lines: the task, the prompt, and whether the result was usable.
- Write down ten tasks from your last two weeks that felt slow or annoying. Emails, reports, planning, research, anything.
- Circle the four that come up most often. Those are your practice material.
If you’ve never opened a chat tool, read our beginner’s guide to AI first. It takes ten minutes and saves you a week of confusion.
Week 1: Writing With AI
Writing is the easiest win, so start there. Every day, take one real piece of writing you need to do and draft it with AI.
Monday might be a follow-up email to a client who hasn’t paid. Tuesday, a short update for your team. Wednesday, a product description or a social post. Give the AI context every time: who it’s for, what you want them to do, and how you usually sound.
Draft a polite but firm email to a client whose invoice is 21 days overdue. We’ve worked together for two years and I want to keep the relationship. Keep it under 120 words. Here’s a past email I wrote so you can match my tone: [paste an email you wrote]
Finish line: five real drafts sent, and a note in your log on which details made the output better. Most people find the same thing. Context and an example of their own writing matter more than clever wording.
Week 2: Thinking And Deciding With AI
Now use AI as a sparring partner. This week you bring it a decision or a problem each day and ask it to push back.
- Ask it to list the risks in a plan you’ve already made.
- Give it two options and ask which one it would pick and why, then argue with it.
- Paste a messy page of notes and ask it to turn them into three clear next steps.
- Ask it what questions you should be asking before you commit.
I’m deciding whether to hire a part-time assistant for 15 hours a week or keep doing the admin myself. Here’s what my week looks like: [paste]. Ask me five questions first, then give me your recommendation and the biggest risk in it.
Finish line: one real decision made with AI’s help, and a note on where it was useful and where it just agreed with you. Learning to spot that second one is a key skill.
Week 3: Research You Can Trust
Week three is about finding and checking information. AI can summarize, compare, and explain. It can also make things up and sound certain while it does it.
Each day, research one real question from your work. Compare two software tools, understand a new rule in your industry, or summarize a long PDF someone sent you. Then check two facts from each answer yourself.
Summarize this 20-page supplier contract in plain English. List every date, fee, and cancellation term in a table, with the page number where each appears. [paste contract text]
Asking for page numbers makes checking fast. Our guide on AI research without hallucinations goes deeper on this. For anything legal, medical, or financial, treat the output as a first read, and have a professional check anything you act on.
Finish line: five research tasks done, with at least one mistake caught and logged. If you didn’t catch one, you weren’t checking hard enough.
Week 4: Build One Repeatable Helper
By now your log shows patterns. Maybe you draft the same kind of email every week, or you always turn meeting notes into a task list. This week you turn your best prompt into something you can reuse in one click.
- Pick the task you did most in weeks one to three.
- Take your best prompt and add everything you had to explain more than once: your role, your audience, your format, your tone.
- Save it where you’ll find it. That could be a text snippet, a pinned note, or a custom assistant if your tool supports one.
- Use it every day this week and edit it each time the output misses.
If your tool lets you build a saved assistant, our guide on building a custom GPT shows the steps. A saved prompt in a notes app works almost as well.
Finish line: one helper you’d miss if it disappeared.
How To Learn AI Beyond Week Four
Look at your log. Count the tasks that are now faster. For most people it’s somewhere between five and ten, and that’s the real measure of whether you’ve learned anything. Then pick the next four tasks from your original list of ten and run the loop again.
This is how to learn AI in a way that sticks. You don’t finish. You keep adding one more task to the pile of things AI helps with, and the pile keeps growing.
Get The Four Foundation Guides In One Place
The Core Four Bundle follows the same order as this plan: getting started, everyday tasks, research and decisions, and prompting properly. Each book has a worked example, a prompt pack, and a cheat sheet you can keep next to your keyboard for the four weeks.
