The Tools and Methods We Use to Teach AI Assistants

When I first started teaching adults about AI assistants, I made the mistake many instructors make: I walked learners through what the tool could do, menu by menu. People nodded politely and then went home unsure what to type. These days the module begins with a task someone already has, such as a letter to a landlord, a week of meal planning or a question about a medical appointment form.

That choice shapes everything else. A learner who has typed a genuine request and read a genuine answer understands the tool far faster than one watching a demonstration. They also notice, without being told, when an answer sounds confident but does not quite fit their situation, which is the most important lesson of the whole module.

Each lesson is short, so no one is asked to hold more than one new idea at a time. We finish with a practice task that can be repeated at home at your own pace.

• Start from a task you already have
• One new idea per lesson
• A practice task after every lesson

Much of our teaching material is a small set of written prompts that learners can copy, change and reuse. They are not clever tricks. They are plain sentences that say who you are, what you need and what form the answer should take, for example asking for a polite email in three short paragraphs or a packing list for a weekend trip.

We give these out on paper as well as on screen, because many of our learners prefer to read away from the laptop and mark up what they want to try. The sheets leave blank lines for your own wording, and that is deliberate. The goal is that you can write your own request within a few weeks, not that you rely on ours.

Over the term we ask learners to rewrite a prompt three times and compare the results. Seeing how a small change in wording changes the answer teaches more than any explanation I could give.

• Plain sentences, not tricks
• Printed sheets with space for your own wording
• Rewrite and compare the results

If I could teach only one thing in this module, it would be to check. AI assistants can write fluent, friendly text that is simply wrong, and they do so without any hint of doubt. So every exercise includes a checking step, where learners compare an answer against a trusted source, such as an official government page, a bank's own website or a family member who knows the subject.

We practise this with questions where the answer is easy to verify, like opening hours, a simple definition or a date. Learners see for themselves that the tool is sometimes right and sometimes not, and that its tone gives no clue which.

This is not meant to make anyone afraid of the tool. A calm habit of checking is what lets people use it with confidence, in the same way you would double-check a phone number you heard second-hand.

• Compare against an official source
• Practise on facts that are easy to verify
• Confidence comes from checking, not from trusting

There are things we do not ask learners to type into an assistant, and we say so in the first lesson. Banking details, health card numbers, passwords, full home addresses and anything about another person's private affairs stay out. Instead we use made-up details in practice, such as a placeholder street or a general description of a situation.

This matters because many people assume a chat window is as private as a conversation with a friend. It is not, and we explain in plain language what happens to what you type and how to find the privacy settings in the tool you choose.

We also keep the exercises away from decisions with serious consequences. An assistant can help you organize questions for a lawyer or a pharmacist, but it should not replace them. Learners leave knowing where the line sits and why.

• No banking details, passwords or health numbers
• Use general descriptions in practice
• Prepare questions for professionals, not replace them

Our method relies as much on pace as on materials. Lessons are recorded in short segments you can pause, replay and return to, and each one uses large, uncluttered screen views so nothing is hard to read. I narrate what I am clicking and why, and I leave silences long enough for you to follow along.

We expect questions that might feel small, like where to find the text box or why the screen looks different from last week. Those are treated as ordinary parts of learning, never as interruptions. Written replies from the instructors aim to be short, concrete and free of jargon.

Learners tell us the relief comes in the second or third week, when the tool stops feeling mysterious and becomes one more thing they can use or set aside as they choose.

• Short, pausable lesson segments
• Large, clear screen views
• Plain-language answers to every question

By the end of the module, most learners can write a clear request, read the response with a sceptical eye and decide whether it is useful. They can draft and tidy a letter, plan a week, ask for a simple explanation of an unfamiliar word and recognize when an answer needs checking elsewhere.

What we do not promise is mastery or any particular result. Tools change, and what you learn here is a way of approaching them rather than a fixed set of button presses. Some people finish and use an assistant daily; others decide it is not for them, and that is a perfectly good outcome.

Either way, they understand what these tools can and cannot do, which is the point of the module. If you have used a computer for email and browsing, you have all the background you need.

• Write clear requests and review the answers
• Know when to check elsewhere
• No prior AI experience required