I’ll be straight with you — I spent the first six months using AI completely wrong. And the embarrassing part? I thought I was being incredibly smart about it.
Every morning I’d open ChatGPT, type something in, grab whatever came out, and move on. Write this email. Summarize that article. Give me an Instagram caption. It worked — technically. The output existed. But something was always off. It was generic. It didn’t sound like me. And honestly, I wasn’t even saving that much time, because I kept rewriting everything anyway.
Six months in, something shifted. I had a client project with a two-day deadline, an unfamiliar topic, and three other projects already demanding my attention. That night, almost out of desperation, I used AI differently — not to do my work, but to think alongside me. The result floored me.
Four hours. The whole thing done in four hours. The client called it the best work I’d delivered. And sitting back after, I finally understood what I’d been missing all along.
This post is about that gap — between using AI and actually using AI well. No theory. No hype. Just the things that changed how I work, tested in real situations, on real deadlines.
The First Mistake: I Was Extracting Output, Not Thinking Together
Most people treat AI like a vending machine. You put in a request, something comes out, you move on. I did exactly this for half a year and wondered why the results always felt hollow.
The shift happened when I stopped asking AI to do things and started asking it to help me think. There’s a meaningful difference between those two.
Here’s a concrete example. I had a difficult client situation — a miscommunication that had been building for weeks. My instinct was to ask AI to write a response for me. Instead, on a whim, I tried this:
I want to explain a situation I’m dealing with. Don’t give me a solution yet — just listen. Tell me where my thinking might have gaps, what I might be missing, and if there’s an angle I’m not seeing.
What came back made me genuinely uncomfortable. Because it was right. It pointed out things I hadn’t wanted to look at — assumptions I was making, ways I was framing the situation that weren’t entirely fair. That discomfort led directly to solving the problem.
From that point on, I started treating AI less like a tool and more like a thinking partner. Someone to pressure-test ideas with. Someone to ask the questions I wasn’t asking myself. The output quality didn’t just improve — the entire nature of what I was getting changed.
The Three Hours I Wasted — and Then Got Back
There was a specific afternoon I think about a lot. I needed a 2,000-word blog post. I typed: “Write a blog post about AI tools.” Three seconds later, a complete article appeared. I thought I was done.
Then I read it.
It was so painfully generic that I actually laughed. No real examples. No point of view. No reason for anyone to read it over the thousand identical articles already online. I deleted the whole thing.
Fresh start — but different approach this time:
I’m writing a blog post for freelancers and small business owners who are curious about AI but overwhelmed. I want to write it myself — give me a rough structure only, no actual writing. Just the skeleton.
I got an outline. Then I wrote each section myself — my examples, my voice, my honest take. When I got stuck somewhere, I’d ask specific questions: “This paragraph feels flat, what’s missing?” or “Can you suggest a stronger way to open this section?” AI became my editor, not my ghostwriter.
That post ended up being one of my most-read pieces that month. Multiple people commented that it felt real. It did — because it was mine. AI shaped it, but it didn’t write it.
The lesson isn’t “don’t use AI to write.” The lesson is: don’t outsource your voice. Use AI to break the blank page, to sharpen your structure, to push back on weak sections. Keep the pen in your hand.
Writing Good Prompts Is a Skill — and Nobody Tells You That
This took me an embarrassingly long time to figure out.
I used to think AI was essentially magic — ask anything, get a good answer. What I slowly learned is that AI is more like a very well-read friend who’s just met you. Ask a vague question, get a best-guess answer. Give them actual context about your situation, your goals, your constraints — and suddenly you’re having a genuinely useful conversation.
Same task, two completely different prompts:
Version one: “Help me write a freelance proposal.”
What I got: a template I could find on any freelancing blog, stuffed with placeholders and hollow professional language.
Version two: “I’m a freelance content writer with two years of experience. I’m pitching to an e-commerce brand that needs 8 blog posts per month. Their main pain point is that their existing content isn’t doing anything for SEO. My advantage is that I understand both SEO and conversion copywriting. The tone should be confident but not stiff — like a real person, not a corporate template.”
What I got: something I could send after twenty minutes of light editing.
The four things I now include in almost every important prompt: who I am in this context, what I’m actually trying to achieve, who the audience is, and what tone or format I’m going for. That’s it. Just those four things transform the output from generic to genuinely useful.
It sounds obvious written out. But most people skip all four and then wonder why AI doesn’t seem to help them.
The Morning Everything Clicked
A few months ago, I had one of those weeks where everything lands at once. Three client deadlines colliding. An urgent revision from a project I thought was finished. A new inquiry that needed a quick response. My to-do list had grown into something I genuinely couldn’t look at without feeling a kind of low-grade panic.
Half the day evaporated just trying to figure out where to start.
That evening, somewhere between frustrated and defeated, I just dumped everything into a chat. All of it — unorganized, no structure, just the mess exactly as it was in my head. Then I wrote: “This is my situation. Help me figure out how to approach today and tomorrow.”
What came back was a clear breakdown — what needed to happen today, what could wait until tomorrow, what was genuinely urgent versus what only felt urgent. Five minutes of work. The rest of the day was the most productive I’d had in weeks.
That became a daily habit. Every morning: brain dump first, structure second. Everything that’s sitting in my head — work, personal, vague anxieties, half-formed ideas — all of it out, into the chat, and then a clear plan back. It works like a personal assistant who never judges, never gets tired, and is always available at 6am.
The difference isn’t that I’m working more hours. It’s that I’m consistently starting with clarity instead of noise.
The Tone Check — Small Habit, Surprisingly Big Impact
I want to talk about something that doesn’t come up in most AI productivity content, because it saved me from at least two professional relationships going sideways.
We all write emails when we’re stressed, tired, or quietly frustrated. We think we sound professional. We often don’t. We sound clipped, or passive-aggressive, or just harder than we meant to be. The email goes out. The dynamic shifts slightly. Over time, those shifts add up.
Now, before I send anything important — a client update, a difficult conversation, any message where the relationship actually matters — I write the draft honestly. Whatever I’m actually feeling goes in first. Then I hand it to AI:
Read this message and tell me how it comes across. Is the tone off anywhere? Does anything sound aggressive, passive-aggressive, or unclear? Would anything here land badly?
Sometimes the feedback is: it’s fine, send it. Sometimes it catches one line that’s sharper than I realized. Sometimes it suggests a complete restructure.
There was a billing dispute with a vendor earlier this year. I was genuinely annoyed and it showed in every sentence. AI flagged it immediately — pointed out three specific lines that would likely escalate rather than resolve the situation. The revised version it suggested was firm, clear, and completely professional. I sent that. The issue was sorted within two hours.
Had I sent the original? I don’t know. Maybe fine, maybe not. But I do know I didn’t have to find out.
A Few More Things Worth Knowing
When you want to learn something new, don’t ask AI to explain it generically. Say: “Explain this like you’re teaching someone who’s smart but completely new to it — use a real, practical example.” The difference between that and a textbook explanation is significant.
For procrastination — and I mean the kind where a task is so big it just sits there for days — try this: ask AI to break the task into steps so small that there’s genuinely no excuse not to start. Each step should take fifteen minutes or less. The first step on the list is usually something like “open a new document and write the title.” That’s it. But once you start, momentum tends to carry you further than you expected.
The weekly review is something I stumbled onto almost by accident and now wouldn’t give up. Every Sunday evening — ten minutes, quick summary of the week. What happened, what didn’t, where time went, what felt good. I give it to AI and ask: “What pattern do you see? Where should I focus differently next week? Give me three specific suggestions.” Two months of doing this and I noticed I’m sharpest on Wednesdays, nearly useless on Monday mornings, and that Friday afternoons are when I make my worst decisions. I restructured my schedule around those facts. The improvement was immediate.
And the last one — probably the most valuable. Before any significant decision, make AI play devil’s advocate. Lay out your plan, then say: “What could go wrong here? What am I not seeing? If this fails, what’s the most likely reason?” It’s uncomfortable. That’s the point. An honest friend doesn’t just tell you your idea is great. AI, prompted correctly, won’t either.
What I Wish I’d Known Six Months Earlier
Not that AI is magic. Not that it solves everything. Not that it replaces the hard parts of thinking.
What I wish I’d known is that AI is a multiplier — but only if you give it something worth multiplying. Your judgment, your voice, your specific experience, your actual point of view: these matter more than the tool. AI makes them faster and sharper, when you use it right. It can’t substitute for them.
For the first six months, I thought I was using AI. I was actually just consuming its output. The difference between those two things is everything.
The real use starts when you think with AI — not just hand it tasks.
So try one thing today. One problem, one honest conversation with AI — not asking for an answer, but thinking it through together. Give it context, give it your actual situation, let it push back.
See what happens.
I have a feeling it’ll be different from whatever you’ve tried before.
More posts like this live at Aiworko — covering AI, productivity, and smarter ways to work. Visit: aiworko.com