AI Wrote Your Marketing and Now You Sound Like Everyone Else

AI wrote your last forty posts. They were clean, on time, and grammatically perfect. And they brought in zero calls.

Not because the writing was bad. Because there was nothing in them that only you could have said. AI writes the average, and the customer picks you for a specific observation that only someone who did the job would know. So don't feed it a style guide. Feed it one real detail from the last job.

I run a door company in Dallas-Fort Worth, and I run Hivelead, where I help owners of $1M to $5M businesses figure out why they stopped growing. This is part of a series on when a business should actually use AI, and it's about marketing, which is the first place most owners point AI and the place it quietly does the most damage.

How it happened

You were supposed to be posting, and you weren't, because writing is slow and you've got a business to run. So you handed it to AI.

Now you post three times a week. The posts look like marketing. They sound like a company. And a customer scrolling past can't tell you from the four other door guys, or plumbers, or roofers, who did the same thing.

Somebody ran an experiment on this. Two thousand AI-written articles, tracked for sixteen months. 1,381 clicks total, less than one click per article. That's not a content problem. That's a sameness problem.

AI writes the average

It's not the tool's fault. AI writes the average. It's built to.

Ask it for a post about why your door sticks in summer, and it gives you the post every door company on earth would get, because it's pulling from everything every door company has ever said. Polished, correct, and identical.

When everyone says the same thing, the customer picks on price

When every business in your category says the same thing, the customer has no reason to pick you. So they pick on price.

You just paid a subscription to turn yourself into a commodity. And since the free channel closed, being indistinguishable costs more than it used to.

Want to see what that's costing you? The Revenue Leak Calculator puts a dollar figure on each of the five places revenue leaves. It's free and there's nothing to sign up for.

Why "feed it your brand voice" isn't the whole fix

The advice you'll read is to feed it your brand voice. Build a style guide. Give it your tone.

You should do that. But for a business your size, it's probably not the right fix, because you don't have a brand voice problem. You have an observation problem.

You have an observation problem

Here's the thing that makes an owner pick you. It's not tone. It's a specific detail about their problem that only someone who's actually done the work would know.

The homeowner reads "your door sticks because the house settled on the north side and the hinge side dropped an eighth of an inch, and here's how you can tell from the gap at the top," and something clicks. That guy has seen my door. He's been in houses like mine. I'll call him.

That sentence can't come from AI, because AI has never been inside a house. It can only come from the job you did on Tuesday.

That's what you fed it. Nothing. You gave it the category and it gave you the category back.

The fix: tell it about the last job

Don't write a style guide. Tell it about the last job.

Every post starts with one real thing you noticed this week. One job, one detail. The door that stuck because of the north side settling. The property manager who called about one door and had six more that didn't latch. The customer who'd been quoted a full replacement when the frame just needed shimming.

Thirty seconds into a voice memo on the way to the next job. Then hand that to the AI and say: write the post around this, don't add anything I didn't say. It's the same method as getting a process out of your head, applied to a post instead of an SOP.

The tool does the boring part

Now the tool is doing what it's good at. Structure, grammar, length, the boring part. And the one thing in the post that matters, the observation, came from you.

The customer reads it and thinks, this guy knows my problem. Nobody else in your category can write that post, because nobody else was on that job. And the commercial customer you want reads it the same way: this is a vendor who notices things.

AI pointed at the category gives you the category. AI pointed at your Tuesday gives you a post only you could have written, in a tenth of the time. It's the marketing version of keeping AI behind the conversation, not in it.

Specific beats polished

A rough post with one real detail in it will out-pull a perfect post with none. So if you only have time for one, skip the polish and keep the detail.

AI didn't make your marketing sound like everyone else. You did, by giving it nothing of yours to work with. Tell it about the last job. Then let it write.

FAQ

Why isn't my AI-generated content getting customers?

Because AI writes the average of everything in your category, so your posts read the same as every competitor who used the same tool. When every business says the same thing, the customer picks on price. The fix isn't better prompts for tone. It's giving the tool one specific observation from a real job that only you could have made.

How do I make AI content sound like me?

Don't start with a style guide. Start with the last job. Record thirty seconds about one real thing you noticed, hand the transcript to the AI, and tell it to write the post around that detail and add nothing you didn't say. The tool handles structure and grammar. The observation is yours, and that's the part the customer responds to.

Does AI-generated content hurt SEO?

It performs poorly when it's generic. One experiment published 2,000 AI-written articles and tracked them for 16 months. They received 1,381 clicks total, less than one click per article. Content built around specific, first-hand observations performs differently because it says something no other page says.

What should a service business post about?

The specific thing you noticed on the last job. The reason a door stuck, the problem a property manager didn't know he had, the repair a customer was overquoted for. One detail per post, in your own words, then let AI do the formatting. Specific beats polished.

Related reading

• Should You Use AI in Your Business?
• Stop Using AI to Talk to Your Customers
• You Don't Have an AI Problem. You Have a Missing-System Problem.
• Why Your Marketing Got More Expensive
• How to Get Higher-Paying Clients Without Becoming a Different Business

Sources

• Search Engine Land, Bogdan Babiak (SE Ranking), March 2026, "How AI-generated content performs in Google Search: A 16-month experiment" (2,000 AI-written articles, 1,381 clicks over 16 months; 244 in the first month)
• Sword and the Script, Frank Strong, March 2026, "Study finds AI-generated content performs poorly in search"