AI for Non-Technical People: Where to Start (Honestly) — Rafaela Buena

If you’re non-technical, the place to start with AI is one tool (ChatGPT or Claude), used on one real task from your actual life, for one week. Not a course. Not a YouTube playlist you’ll save and never open. You open the tool, you bring it a real problem, and you talk to it in plain sentences, because plain sentences are literally the interface. That’s the whole entrance exam.

I know it doesn’t feel that simple, so let’s take the feeling apart.

First, the permission slip

I am not an engineer. I have never written code the traditional way. And I built Gut Goddess, a whole gut health app, real users and everything, using AI and no technical team. Along the way I once spent a confused half hour on a “broken” feature before noticing my caps-lock was on. I am telling you this so you understand exactly who is allowed to use AI: silly humans. All of us.

The tech industry has spent decades training you to believe that anything powerful requires a translator. AI is the first genuinely powerful tool where the translator is built in. You type words. It types words back. Your forty years of experience talking to people is, for once, the relevant qualification.

The “non-technical” label mostly measures who felt welcome, not who is capable. You’re capable.

Your first week, concretely

Pick ChatGPT or Claude. Free version is fine. Flip a coin if you must; the choice matters far less than starting.

Then, for one week, bring it one real thing a day:

A tricky email you’ve been avoiding. Paste it, explain the situation and what you want, ask for three reply options.
Something you don’t understand. A contract clause, a medical letter, your own industry’s latest buzzword. Ask it to explain like you’re smart but busy, because you are.
A decision you’re chewing on. Describe it, ask what you might be missing, argue back when you disagree. Especially argue back.
Dinner. Genuinely. “I have these seven ingredients and twenty minutes.” Low stakes, instant payoff, and it teaches you the rhythm of asking, getting, and refining.

Notice the pattern: your life, not tutorials. The people who stall out are almost always the ones practicing on pretend tasks. Fake tasks produce fake learning.

None of this has to be solo. The quickest version is doing it on your own real work with someone beside you, which is exactly what happens in an AI workshop for your team. Prefer to work through it privately, at your own pace? That is what 1-1 AI coaching is for.

What you can safely ignore

The pressure to keep up with every model release. The bros announcing weekly that everything just changed forever. Prompt marketplaces selling you “1,000 proven prompts” as if good conversation came in bulk. Courses with countdown timers. Anything that makes you feel behind, because the secret is that almost everyone is at the beginning, including many people who talk loudly at dinner parties as if they aren’t.

You can also ignore, for now, the technical vocabulary. Tokens, parameters, fine-tuning.. lovely words, entirely optional. You don’t study carburetors to drive to the beach.

One thing you may not ignore: AI confidently makes things up. It will state a wrong fact with the serene authority of a museum plaque. So the one rule you adopt from day one is simple. Anything factual that matters gets checked before you rely on it or repeat it. That single habit puts you ahead of a shocking number of loud experts.

The shift that actually matters

Somewhere in week two or three, something quietly rewires. You stop asking “what can AI do?” and start asking “could AI do this?” about your own tasks, unprompted, mid-annoyance. That question, arriving on its own, is the actual milestone. Not a certificate. A reflex.

From there it compounds. The email trick becomes a reporting trick becomes, who knows, an app. Mine started with a question I typed on another continent, under a Ghanaian ceiling fan, about my own confusing gut. No engineering degree ever entered the story.

FAQ

Which AI tool is best for beginners?

ChatGPT or Claude, and honestly either one. Both understand plain language, both have free tiers, both forgive clumsy questions. Beginners lose more time comparing tools than they’d lose picking the “wrong” one, because for everyday tasks there is no wrong one.

Do I need to learn prompt engineering first?

No. Start by writing to the AI the way you’d brief a smart new assistant: context, what you want, what you’ll do with it. That’s most of prompt engineering anyway. Technique grows naturally out of use, not the other way around.

How long does it take a non-technical person to get comfortable with AI?

About a week of daily use for basic comfort, a month for it to become reflex. The variable isn’t intelligence or age, it’s whether you practice on real tasks from your own life instead of watching other people use it.

And if you’d rather not start alone, don’t. Sitting beside someone for their first real session, watching the “oh, THAT’S it?” land on their face, is my favourite part of this job. Book a 1-1 AI session through my contact page, or bring your whole wonderfully non-technical team for a workshop. Bring the task you’ve been avoiding. Especially that one.

Keep reading

What is prompt engineering (and do you actually need it in 2026)?

AI for non-technical people: where to start (a gentle, honest guide)