AI can write a blog post, a caption and an ad in seconds. That's exactly why so much of it gets ignored. This week, several of the marketers we follow shared data and opinions on what happens when everyone uses the same tools, and what still gets through.
Here's what they said, and what it means for a Las Vegas business owner deciding how to use AI in their marketing.
AI is doing half the publishing and little of the work
Neil Patel shared numbers from his team's study of 100 companies. AI-generated content made up 52.1% of what they published but brought in just 4.9% of their organic traffic. Human-made content was only 14.5% of what they published and brought in 87% of their traffic.
He also cited a Graphite analysis of 55,400 articles that found 49.9% of new articles were mostly written by AI, while 86% of the articles showing up in Google search were human-written.
In a 744-article test across 68 websites tracked for 12 months, Patel says human-written articles were getting 5.4 times more traffic per post by month five, and that gap held through month 12.
Platforms are filtering generic content
Patel listed several platform changes: Google's 2024 policy on scaled content abuse, LinkedIn pulling generic AI posts out of recommendations, a LinkedIn button to report AI slop, and Snap no longer recommending fully AI-generated videos on Spotlight. He said LinkedIn described the posts it holds back as content that looks AI-generated "and has no clear point of view."
He also warned about a cost most owners don't see: publish 50 generic pages on one topic and they compete with each other, which can drag down a good page you already had.
Everyone gets the same ideas
Codie Sanchez put it bluntly: "Average and mediocre is now the cheapest thing you can be because a machine makes it for free. Original is expensive again because it's scarce again." She cited research from UCLA and Exeter finding that writers who used AI ideas scored higher individually, but their stories collectively sounded the same.
Kieran Flanagan of Marketing Against the Grain went further, saying he's coming to believe "the best way to grow an audience and to create great content is to use no AI at all. Not even for ideation," because prompts tend to lead everyone to similar ideas and research. In another episode, AI educator Ruben Hassid compared writing to flirting: "if everyone is using the same pickup lines... it becomes boring." He also said "writing is thinking," and readers can tell when someone never lived what they wrote about.
The counterpoint: volume still matters
Not everyone agrees on how much to publish. GaryVee said this week that "quantity is not debatable... quality is subjective," and that he believes "quantity is a meaningful competitive advantage." His advice to every business is to start making content now, whether by interviewing people, ranting or documenting your day.
Patel's advice was nearly the opposite: "You're probably better off publishing less." The way to square these: volume of real material from your business is good. Volume of generic material from a prompt is what gets filtered.
The test every post should pass
Two creators offered almost the same test this week.
Neil Patel: read your draft and ask whether a competitor with the same AI could have published it. "If the answer is yes, it's slop, no matter how well it's written."
Sabri Suby: ask yourself, "could I get this information from ChatGPT or Claude... If so, then you're making the wrong type of content." His alternative is to go deep on what only your vantage point can show.
What only a local business can publish
Patel's examples of content AI can't copy: your own data, a real result with a number, a customer story with a name on it (with permission), or a test that failed and what you learned. For a Las Vegas business, that might look like:
- What you found inside a 15-year-old AC unit in a Henderson attic last July
- The three questions every new moving customer asks, and your honest answers
- A before-and-after job with the actual timeline and what went wrong
- A short video of your crew explaining a repair in your own words
None of that can be generated from a prompt, because it only happened at your business.
Where AI still helps
None of these creators said to ban AI from your business. Patel said he uses it every day for research, outlines, transcribing interviews and cleaning up grammar "after you've written the thing." Hassid uses it heavily for routine emails. The line they draw is the same: let AI handle the busywork, and keep the part where you say something only you could say.
A weekly rhythm that works
Patel's suggested week: on Monday, write down one thing that actually happened in your business. On Wednesday, write it up yourself with a number in it. On Friday, publish it. Then measure replies, inquiries and time on page instead of impressions.
At TAAM, that's how we approach content for local businesses: film the real work, tell the real stories, and use AI only to speed up the parts customers never see. If you want to see what that looks like for your business, start with a free week of video.

