Look – we’re drowning in content on content. Every day there’s another “definitive guide” or “ultimate framework” for writing with AI and honestly? Most of it’s making things worse, not better.
And I’ve found that the more advice we read about AI writing, the more paralyzed we become. It’s like . . . you know when you are standing in front of your closet with a hundred outfits but just can’t figure what to wear? That’s us with AI prompts. We have all these complicated strategies, these elaborate workflows that look great on paper, but then when you actually try to make something happen, we stumble.
There is this funny thing that our brains do. We assume big problems require big and complex solutions. So when somebody says “hey, try this simple thing,” we shrug it off almost immediately. Too easy. Can’t possibly work. If it was that easy, wouldn’t everybody be doing it already?
But this is what makes these shortcuts secret. So powerful, they’re hiding in plain sight as they are too common to be powerful.
We use AI for so many different things that I would like to share some tricks I use that radically altered my work to AI that are so obvious you may roll your eyes. But stay with me, simple does not mean weak. Sometimes the best tools are the ones we keep stumbling into while we are looking for something more complicated.
Stop Starting from Scratch (Please, Just Stop)
This is going to sound counter-intuitive to start with, but bear with me. Most people – and I did this for months too – open up their AI tool and basically say “write me content about X.” Then they question why everything sounds as if it was written by a very polite robot who took a community college writing course.
The missing secret that everyone is missing? But first, teach AI how to do good. Don’t start from nothing, start from something great.
I know what you are thinking, “If I already have great content, why do I need AI?” But you’re not copying (that would be… problematic – legally and ethically). You’re teaching. There’s a massive difference.
Here’s what actually works: Read an article you absolutely love, something that made you think “damn, I wish I’d written that.
” May be from your industry, may be from somewhere totally different. Paste that in your AI chat and tell it something like, “This tone, this structure, this depth–I want you to make something with these characteristics, but about [whatever your topic is].”
It’s kind of like… say you’re teaching somebody how to cook. You would not merely give a definition of what “delicious” tasted like. You’d let them taste something good, and teach them to make it. AI is no different – great at pattern matching, awful at interpreting abstractions.
A friend of mine (she’s a freelance marketer, works mostly with B2B clients) was pulling her hair out trying to write case studies. She’d spend literally hours adjusting prompts, tweaking parameters, testing different AI tools. Nothing clicked. The content was always flat, generic and it lacked that something that makes you actually want to keep reading.
Then she tried this approach. Found a case study of a competitor that was absolutely stellar – the kind that tells a story, instead of just listing features and results. She showed it to Claude and said, “You have to beat this energy but for my client’s story.”
First draft? Her client agreed to it with like, three minor tweaks. That was something that had never happened before. Ever. She told me she came close to crying (okay, she cried a little bit, but don’t tell her I told you that). Now she cranks out case studies in less than an hour instead of killing a whole afternoon to the content gods.
The pattern is more important than what is said. AI is almost too easy to use – give it a great pattern, get a great result.
Talk to It Like a Person and Not a Vending Machine
This is a big one and it’s where most people completely fall down. We use AI just like a vending machine. Put prompt in, get body, go on about your day. Transaction complete. But that’s not how good writing comes about — not for humans, not for AI, not for anything.
The real shortcut? Have an actual conversation. Go back and forth. Iterate. Push back. Ask questions.
People do not do this because (and I get it, time is money) they want to see results right away. One prompt, done. But here’s the thing – that transactional mindset generates transactional content. Top of page content that people scan and forget before they’ve even scrolled through it.
What does work: Start with something vague, even messy. Something like “I have to write about AI content strategies but I don’t know where to even begin.” Let the AI respond. Then, and this is important, don’t simply take what’s offered to you. Engage with it.
Questions such as, “Could you say the second point another way?” or “What objections can be made against this?” This is too formal, can we make it more conversational? Each exchange provides layers, nuance, texture that cannot be achieved in one shot.
There is this B2B SaaS company (I can’t remember the name, but they are in the project management space) that was writing product comparison articles. Their approach initially was these massive 300 word prompts that specified everything – tone, structure, key points, SEO keywords, the works. Very thorough, very organized.
Content was accurate but… God, it was boring. Like reading a technical book by someone who’d never enjoyed reading.
When they switched to this conversational style of writing, by starting small and then iterating with questions like “how about making it more specific,” “fill in some actual data here,” and “assuming I’m not going to close the tab immediately, please rewrite this intro,” their engagement numbers doubled. Within a month. Doubled.
And here’s the bonus (the part that nobody mentions): each conversation teaches you what works. You are not just getting better content, but you are getting better at prompting. It compounds over time.
Your Voice Aren’t Descriptors – It’s Samples
Every one of the guides tells you to “specify your tone.” Use a “friendly and professional” style. Make it a way with them but make it authoritative. And you know what happens? Everything sounds the same. That general AI voice that is… OK. Competent. Completely forgettable.
The hidden secret that makes all the difference? Don’t describe your voice. Show it.
People believe (I thought this too, honestly) that voice of AI should be able to get out of the descriptions alone. Like some sort of writing mind-reader. But it can’t. Descriptions are too vague, too open to interpretation.
What AI can do – and this is almost eerie the first time you see it – is to replicate real-world cases with uncanny precision.
Take three to five pieces of your writing. Could be emails you’ve written, social media posts, past articles, even text messages if they are substantial. Git them in your chat and say: “Here is how I write: This is my voice. Write a paper about [whatever your topic is] following this pattern.
The transformation is… For that, I don’t have a good analogy for it. It simply works in such a magical way until you know what’s going on behind the scenes.
I was talking to a newsletter writer the other day (she has around 15k subscribers, writes about remote work and productivity). She was wasting three, four hours editing AI drafts to sound like herself. Three hours. On editing. Which type is totally counterproductive to the point of using AI, right?
Then, she condensed her favorite newsletter pieces – the ones where people responded back saying ‘this sounds exactly like something I needed to hear today’, and used these as her voice template.
The AI produced one draft which required twenty minutes of tweaking.
Twenty minutes instead of three hours plus. And–this is the part that convinced her that it really worked–her audience didn’t notice any difference. The authenticity of it remained intact, the engagement didn’t go down, people continued to respond with their personal stories.
Your voice is not “friendly” or “professional” or “witty.” It’s patterns. Pattern of words, rhythm patterns in sentences, way of tying ideas together. AI is nothing but a pattern matching machine on steroids. Try giving your actual patterns rather than abstract descriptions, and see what happens.
Let AI do the Grunt Work, You do the Human Steps
This may be the most controversial shortcut, since this sounds like me saying “don’t use AI” (I’m not saying don’t, stay with me).
Here it is: Don’t let AI Write Everything. Use it as structure, research, and drafting of the expository sections–but compose the introduction. In short, the conclusion should be written by you. Write any personal experience or special points by yourself.
People ignore this right away because it appears to defeat the purpose. So why use AI when I’m still writing? But this hybrid model, this is where the magic lie.
Think of it this way (and I’m going to need another cooking analogy in here because apparently my brain today is just cooking minded): AI is like a super efficient sous chef. It has the ability to chop vegetables, measure ingredients, follow recipes perfectly. But the creative choices, the seasoning tweaks, the plating, the special technique that makes the dish yours, that requires the head chef.
AI is phenomenal in organizing information. It can create outlines that make sense, create first drafts of explanatory content, research topics and synthesize information faster than any human. But you? Your ability to emotionally bond is enhanced. At offering your own unique insight. At writing those first lines that make people want to keep reading.
So if you think about ways to combine these strengths instead of either trying to make AI do everything or being obstinate and doing everything yourself, then you’re going to get content which is both efficiently produced but also engaging and worthy of the content creator and their brand.
There’s this corporate blogger that I was in touch with last year (works for a fintech company, makes educational content about, well, pretty dry topics like compliance and regulations.) She approached this method almost by accident: she had grown frustrated with the formulaic AI intros, so she began writing them herself.
Now there’s her process: AI comes up with the outline, drafts the middle sections of the document (the how-to parts, the explanations, the data). But she writes her own intros with personal anecdotes, and her own conclusions with specific, actionable advice with her actual experience.
Her output increased by 60% — she is producing more content in less time. But the crazy thing is, her average time-on-page also increased by 40%. The content became more interesting and was created faster. That’s not supposed to happen. Usually it’s a tradeoff.
But it does make sense when you think about it. The AI takes on the time consuming structural work, freeing her up to put more energy on the parts that make the connection – the parts that make people actually care about fintech compliance (which, let’s be honest, is a heavy lift).
So… What Now?
These shortcuts work because they’re in line with reality, rather than the way we assume that AI should work.
And yeah, they’re simple. Maybe too easy for some people to take seriously. But simple does not mean weak – sometimes simple means you’ve removed all the unnecessary parts until there is nothing unnecessary left.
Here is my challenge to you (literally, do this): Take one of these shortcuts. Just one. Use this for whatever piece of content you are writing next. Could be tomorrow, could be this afternoon if you’re feeling up to it.
Notice what’s different. Pay attention to what is changing – not just in the output, but in the experience of the process. Is it easier? Faster? Does the content come across as more like you?
Then return and have another go at another one. Stack them. I am still just now learning how to combine them in ways that I haven’t even come close to figuring out (so if you find something, I mean it, tell me–I’m always learning).
The people that are actually succeeding with AI content aren’t using secret advanced techniques that cost thousands of dollars. They’re not quick-witted engineering whizzes with computer science degrees. They’re just regularly putting to use these unappreciated basics when everyone is chasing complexity, buying courses, trying to find the “perfect” framework.
Your competitive advantage isn’t lurking somewhere in the future in AI tool or prompting library. It’s here — and it’s in these simple shortcuts you’ve likely been walking right past because they were too common to seem important.
Time to stop overlooking them, don’t you think?
And if you try one of these and it doesn’t work—or if it works way better than expected—I genuinely want to hear about it. We’re all figuring this out together, and the best insights come from actually using this stuff, not just theorizing about it.
