If every competitor can publish polished posts, polish stops being an advantage. The new episode of ZilckSound, "Everyone's Content Got Good. Now Nobody Is." (Ep. 22, 8:57), argues that the advantage has moved from competence to evidence. Listen to the episode here.
The Problem With "Good Enough"
The episode argues that AI made everyone's marketing content competent this year, and that this is the problem. The risk isn't bad writing. It's that your clear, well-structured post says what your competitor's clear, well-structured post says.
For anyone running a content calendar, this changes the productivity math. Output is cheap now, and cheap output loses value as everyone adds to it. Hours saved on drafting only pay off if you spend some of them on something a model can't produce.
What the Data Says
The episode cites a study of a million pages that found fully AI-written content ranking number one. That matches Ahrefs' June 2026 analysis of 1,000,000 pages from the top 10 results for 100,000 searches. It found that 5.3% of top-3 pages are 100% AI-generated, so fully AI pages can and do rank at the very top.
The same research shows the other side. Pages with under 50% AI content hold 82.2% of top-3 rankings, and 54.7% of top-ranking pages have less than 20% AI content. Ahrefs also notes that heavier AI use correlates with lower rankings, impressions, and indexation. An earlier Ahrefs study of 600,000 pages found almost no correlation (0.011) between AI use and ranking, and concluded that Google neither rewards nor penalizes AI content as such.searchenginejournal+1
My reading is that AI isn't the risk, and it's not a free pass either. Search engines can rank your page, but they can't make it matter to anyone. Sameness is the risk the episode names.
Two Small Businesses, Two Outcomes
The episode follows two owners. A remodeler's polished posts became interchangeable with his competitors'. His content wasn't wrong, but nothing in it could only have come from him. A florist takes the opposite approach and feeds the machine things only she knows.
I'm working only from the episode summary, so I won't guess at her specifics. The contrast still carries a practical lesson: the model supplies the fluency, and you supply the raw material. Your jobs, your mistakes, your customers' questions and your pricing logic are inputs no competitor can copy.
Why Evidence Is the Moat
Google's own guidance points the same way. Its documentation asks whether content provides original information, reporting, research, or analysis. It also describes "effort" and "originality" as quality signals. Originality means offering information or perspectives not already available on other sites, and high effort includes original analysis. By contrast, using generative AI to produce large amounts of text without manual oversight or curation counts as little to no effort.
Google's guidance also says that crediting other sources doesn't replace the need for original effort. Summarizing what's already out there, however well, doesn't count. Practitioners call the idea "information gain": adding something new to the conversation instead of nodding along.
A Practical Playbook
Here are five ways to turn "evidence over competence" into this week's workflow.
- Audit for interchangeability. Cover your logo and read your last five posts. If a competitor could have published them, mark them for rework.
- Build a proof bank. Collect photos, project notes, customer questions, numbers, and failures in one place. Feed these to your AI tools instead of asking it for generalities.
- Add one original element per piece. That could be a data point, a first-person story, a screenshot, or a contrarian opinion you can defend.
- Keep a human editor in the loop. Use AI for outlines, structure, and cleanup, and keep your judgment on claims and voice.
- Measure differently. Track replies, saves, branded searches, and sales conversations alongside traffic.
The Takeaway
Fluency used to signal effort, and it doesn't anymore. When everyone can sound professional, readers and search engines reward the person who shows proof. Press play on Episode 22, then open your last blog post and ask what in it only you could have written.
Listen now: ZilckSound, Ep. 22
Discussion