The “should we even use AI in writing?” debate is quietly dying. The dust is settling, and sure, you’ll still find toxic AI slap-fights on LinkedIn, but those feel more like rearguard battles at this point.
My sense is that most organizations are fine with AI being part of the writing process, as long as the copywriter, content marketer, or content strategist is transparent about it. The question that matters more today is: what do you do after the text is on the page?
AI writes fluently, in a structured way, and with correct grammar. And the quality keeps getting better with every new model. The problem is that “fluent and correct” isn’t always the same as “good.” Publishing a piece from the first moment it looks professional is a genuinely bad idea.
Maybe you only use AI for your first draft. Maybe only for a quality check afterward. Or maybe for the entire writing process from start to finish. However you use AI, there’s still some thinking to do before you hit publish.
Here are five questions you should ask yourself before an AI-generated text goes out the door.
1. Do I actually understand this myself?
Maybe the most obvious question. Not “does it sound logical?” but: can you explain it to your colleague, your mother, or your dog without having the text in front of you?
AI is very good at wrapping an idea in convincing sentences, even when the idea itself is hollow. If you don’t understand the topic more deeply than the text does, you haven’t written an article. You’ve published something you can’t back up.
2. Have I checked everything?
Every number, every claim, every name. If AI says “According to McKinsey, 73% of B2B marketing managers are vegetarian,” you’d better have the actual report from a reliable source in hand (preferably McKinsey’s own site) and have verified that number yourself.
AI can be convincingly wrong. It writes facts and fabrications in exactly the same tone. That’s the dangerous part: it gives you no signal when the facts are shaky. Fact-checking isn’t optional, and it’s definitely not something you squeeze in at the last minute.
3. Would I actually say it this way?
Does it sound like you, or like a corporate version of you? Would you tell the story this way if you were explaining it to the above-mentioned colleague/mother/dog?
AI writes toward the average: correct sentences, neutral tone, no sharp edges. Which is exactly the opposite of what makes your brand distinctive. If you’d read the text out loud and it feels strange in your mouth, that’s your sign you’re not done yet.
4. Did I challenge myself enough on the premise?
This is the question most people skip. AI is excellent at developing an idea, but weak at questioning it. It takes your first angle and turns it into something coherent, even if that angle is the most obvious one, the safe one, the version ten competitors have already written.
A piece can be perfectly structured and still add nothing. So ask yourself: does this article bring enough value? Does it make someone pause, or does it just confirm what they already thought?
5. Does it hold up without the AI gloss?
AI-generated text has its own recognizable fingerprint. Em dashes are maybe the most obvious red flag (although I have nothing against em dashes themselves — that’s a different story), but there are plenty more tell-tale words, sentence patterns, and phrases that reek of AI. “It’s not x, it’s y!” The short sentences built for drama. You probably have your own personal list of AI annoyances.
Strip away those AI tics and see what’s left. Often it’s surprisingly little. A genuinely good piece doesn’t lose substance when you scrape off the slick AI coating. If yours does, the skeleton probably wasn’t strong enough to begin with.
For pieces that need to stand out, I set a high bar for myself and for the AI helping me get there.
AI-generated copy: yay or nay?
My relationship with AI is complicated. I’ve been writing copy for B2B and tech companies for twenty years now, and I still love doing it. Twenty years is a decent amount of experience and pattern recognition to build up. You could think that would make me dismissive of AI-generated text. Yet, I’m not too proud to admit that AI-generated text is often really good. Maybe AI copy is even good enough for 80% of corporate content out there. For that other 20%, the pieces that need to stand out, I set a high bar for myself and for the AI helping me get there. No way am I just passing along whatever the AI hands me.
Sometimes AI is my brainstorming partner, sometimes it gives me a quick outline, sometimes even a first draft. But if I don’t put my own thinking to work too, I don’t dare send my AI-assisted text out into the world. The five simple questions above are how I make sure I do.
None of these questions are about whether AI is “allowed.” What matters for me is: do I stand behind what gets published under my name? How much time and energy am I willing to put in to make sure that’s true? How critical is this piece of writing?
AI may change how the text comes together, but it doesn’t change who’s responsible for what ends up on the page.