If You Are Not Convinced Yet
Three free experiments you can try in ten minutes, no signup required.
If you have landed here first, good. You do not need to have read anything else in this guide. This page stands on its own.
Let me be upfront: I am not an AI cheerleader. There are things about this technology that genuinely give me pause -- the way it can generate convincing nonsense, the privacy implications, the effect it is already having on creative industries, the companies racing to ship products before they fully understand what they have built. I have written about those concerns elsewhere in this guide, and I mean every word.
But I am also a realist and a pragmatist. Some of these tools are genuinely, practically useful -- not in a "wow, look at this party trick" way, but in a "this just saved me an hour on a Tuesday afternoon" way. And here is the thing that changed my mind about writing this guide: AI is not waiting for any of us to be comfortable with it. It is already embedded in search results, email, office software, and the phones in our pockets. The choice is not whether to engage with it. The choice is whether you understand it enough to use it on your terms, or whether it just happens to you.
I would rather you master these tools before they master you. That starts with seeing what they actually do, not what the headlines say they do.
Below are six experiments. They start simple -- rewriting an email, summarising a document -- and build to things that tend to surprise even the sceptics, like turning a phone photo into useful information or building a working app by describing it in English. Each one is free. None require signup or a credit card. You just need a browser and ten minutes.
I would encourage you to work through them in order, because each one builds on the last. But if the first three feel too basic, jump straight to the three that convert the harder-to-impress crowd.
If none of these are useful to you, AI is not for you yet. Check back in six months. No hard feelings.
Experiment 1: Fix a Clunky Email
Time: 5 minutes
What to do: Find an email you sent recently that took you too long to write, or one where you were not happy with how it turned out. Copy the text. Open Claude or ChatGPT and paste it with this prompt:
What you should get back: A tighter, clearer version of your email that says the same thing in fewer words. You will probably find that the rewrite is close to what you wanted to say in the first place but could not quite get there.
Time saved versus doing it yourself: Most people spend 10-20 minutes agonising over an important email. This gets you to a solid draft in under a minute. You will still want to read it and adjust the tone, but the heavy lifting is done.
Experiment 2: Make Sense of Something Boring
Time: 5 minutes
What to do: Find a document you have been avoiding. An insurance policy renewal, a body corporate circular, a terms-and-conditions update, a long report you were sent, anything dense and tedious. Upload the PDF or photograph it with your phone and paste the image into Claude or ChatGPT with this prompt:
What you should get back: A clear summary that would have taken you 15-30 minutes of careful reading to extract yourself. The AI will flag anything important buried in the fine print.
Time saved versus doing it yourself: Depends on the document, but for a 10-page insurance renewal or strata notice, you are saving 15-30 minutes of reading and probably understanding it better because the AI strips out the jargon.
Experiment 3: Plan Something Specific
Time: 10 minutes
What to do: Open Claude or ChatGPT and describe an outing, a dinner, or a weekend plan with real constraints. Be specific. Here is an example prompt, but replace it with your own situation:
What you should get back: A specific, practical plan with real places you can actually visit. Not a vague "why not try a nice walk" suggestion, but an actual itinerary with names and times.
Time saved versus doing it yourself: This replaces the 20-minute cycle of Googling, reading reviews, checking menus, and going back and forth with your partner about what to do. The AI does the legwork. You make the final call.
Still Not Convinced? Three More for the Harder to Impress
If the three experiments above felt too basic, these are for you. Each one demonstrates a capability that genuinely surprises people who thought they already understood what AI could do.
Experiment 4: Argue Against Yourself
Time: 10 minutes
What to do: Think of a position you hold strongly -- a business strategy you believe in, a political view, a life decision you have made, a hiring philosophy, anything you are confident about. Open Claude or ChatGPT and type:
What you should get back: Three genuinely challenging counterarguments that you probably have not considered, or have considered but not in this formulation. The AI is not smarter than you. But it has access to every perspective ever written about your topic, and it has no ego invested in agreeing with you.
Why this matters: This is not about changing your mind. It is about stress-testing your thinking. The best strategists, lawyers, and leaders do this instinctively -- they argue against their own position to find its weaknesses before someone else does. AI makes this available to everyone. If you make decisions for a living, this alone justifies learning these tools.
Experiment 5: Turn a Photo Into Actionable Information
Time: 5 minutes
What to do: Open Claude or ChatGPT on your phone. Walk to your medicine cabinet, your pantry, your filing cabinet, or your garage shelf. Take a photo of something with lots of small text that you would normally never bother reading -- a medication label, a nutrition panel, a warranty document, an appliance data plate, a bag of fertiliser. Upload the photo and type:
What you should get back: A detailed, plain-English explanation of everything in the image -- dosage information, chemical composition, safety warnings, expiry dates, compatibility notes, whatever is relevant. For a medication label, it will explain what the medication does, common side effects, and what to avoid. For a fertiliser bag, it will explain the NPK ratio and what it means for your garden.
Why this matters: You have been walking past unread labels your entire life. Your phone camera plus AI turns every label, notice, and fine-print document into something you actually understand. Once you try this, you will do it constantly.
Experiment 6: Build Something That Works in Fifteen Minutes
Time: 15 minutes (but you will want to spend longer)
What to do: Go to bolt.new (free account). Think of a simple tool that would be useful to you personally. Not something grand -- something specific. A tip calculator that splits bills unevenly based on who ordered what. A packing checklist generator for different types of trips. A wine tasting notes journal. A countdown timer for multiple things at once. A reading log that tracks pages per day.
Type a description of what you want. Be specific about features and layout, just as you would if you were briefing a human developer. Watch it build a working application in real time. Then tell it what to change.
What you should get back: A working, deployable web application that you built by describing it in English. No code written. No developer hired. No technical knowledge required.
Why this matters: This is the experiment that converts the hardest sceptics. Not because a reading tracker is life-changing, but because of what it implies. If you can build this in fifteen minutes, what else can you build? The answer, as the advanced projects section shows, is almost anything you can clearly describe. The gap between "I have an idea" and "I have a working tool" has collapsed, and most people have not noticed yet.
What Just Happened
If any of those experiments produced something useful, you have just seen the core value of AI tools. Not replacing your thinking, but doing the tedious parts faster so you can focus on the parts that matter.
The email experiment shows AI as an editor. The document experiment shows AI as a reader. The planning experiment shows AI as a research assistant. The argue-against-yourself experiment shows AI as a thinking partner. The photo experiment shows AI as a translator of the physical world. The build-something experiment shows AI as a creator. Every section in this guide builds on one of those capabilities.
If all six felt like a waste of time, that is fine too. These tools are not for everyone, and not everything benefits from AI. But most people who try these experiments find at least one of them genuinely useful. And that one is usually enough to make the rest of the guide worth exploring.
Where to Go Next
If you want to understand what you just used, start with What AI Actually Is. If you want to learn how to get better results, go to How to Talk to AI. If you want to know what is safe and what is not, read Privacy and Security. Or if you just want to see more of what these tools can do, browse Writing and Brainstorming or Research -- the two areas where most people get the most immediate value.