I once watched a business owner generate forty social posts in an afternoon using AI. Forty. She was thrilled, practically vibrating. Three weeks later she called me because engagement had cratered and she couldn’t figure out why, and I could hear in her voice she was bracing for me to say “I told you so.”
I didn’t, because I hadn’t told her anything yet, but we went through the forty posts together anyway. Every single one was fine. Grammatically correct, on topic, nothing technically wrong with any of them. But nothing in any of them sounded like her, and her audience, who’d been following her for two years, noticed before she did. Turns out two years of trust doesn’t survive forty posts that could’ve come from anyone.
AI can write faster than you. It can’t decide what’s true for your brand.
Speed is not the problem, and I want to be clear about that, since a sharp reader pushed back on me recently, rightly, that speed itself can absolutely be a strategy. Being first to market, first to test an idea, first to respond when something breaks, that’s real and it matters, and I was wrong to imply otherwise.
The problem is a specific kind of speed: the kind that skips the decision about what’s actually true for your brand and replaces it with whatever’s fastest to generate. Those are not the same thing, and mixing them up is where the forty-posts-in-an-afternoon story comes from.
AI is extremely good at producing content that meets a bar. Grammatically sound, on topic, technically correct, the marketing equivalent of a firm handshake. What it can’t do is know, on its own, that your brand doesn’t do exclamation points, or that you’d never use the word “hustle” un-ironically, or that your actual customers respond better to a specific, slightly self-deprecating tone than a polished one. Those are judgment calls. You either make them, or you let volume make them for you by default, and volume’s default is almost always the most generic version of correct, which is a very efficient way to become forgettable.
Here’s the trap underneath the trap, and it’s the part almost nobody catches until it’s already happened. When output is slow and effortful, you naturally scrutinize it, because you spent real time on it and you want it to count. When output is instant and free, that scrutiny quietly evaporates. Nobody decides to stop caring about quality. It just erodes, one “eh, good enough, next” at a time, because the cost of producing the thing dropped to almost nothing and your attention followed it down. The forty posts weren’t a decision to lower the bar. They were forty tiny moments where the bar was never raised in the first place, because raising it would have slowed down a process that felt good precisely because it was fast.
Here’s the practical test I use with clients: after AI drafts something, read it and ask “would this specific sentence surprise someone if they knew it came from an AI with no brand guidance?” If the answer is no, if it could have been written for literally any business in your category, that’s the signal to go back in and make an actual decision, not just hit approve because it’s technically fine and you have fourteen other things to do.
This is what “wisely, not widely” actually means in practice. Not less AI. Not slower AI. AI aimed carefully at the parts of the job that benefit from volume and speed, with a human decision still sitting in the loop for the parts that don’t. The skill worth building isn’t typing faster or generating more. It’s knowing which parts of the work are safe to hand over and which parts are the whole reason someone chooses you instead of the competitor two blocks down, and guarding those on purpose.
Takeaway: The question isn’t whether AI is fast enough, it’s whether you’re still the one deciding what’s actually true for your brand.
Try this: Pull up the last five things AI helped you write. Run each one through the surprise test above. If more than two or three would pass for “any business in this category,” that’s not a content problem, that’s a missing-decision problem. Pick the worst offender and rewrite just one line of it, the line only your business would say, and notice how much that single line changes the whole piece.
Next issue, the flip side of all this: using AI to actually get more creative, not just faster. Talk soon, Tiffany