Why AI-generated content fails to convert or connect, and what the 'human layer' looks like in practice: opinion, experience, specificity. A practical guide to better content.
We tested every AI content tool available. Most of them. Over several months.
They all produced the same thing: beige nothing. Competent, inoffensive, structurally correct — and completely indistinguishable from every other piece of content in the world.
It wasn't wrong, exactly. It just wasn't anything. It wouldn't win a client. It wouldn't stop a scroll. It wouldn't make a prospect feel like someone understood their problem.
Here's why — and more importantly, what to do instead.
The Technical Reason AI Content Feels Generic
AI language models are trained to predict the most statistically likely continuation of any given text. In practice, this means they produce the average of everything they've been trained on.
When you ask an AI to write a blog post about branding, it produces a blog post that reflects the aggregate of every branding article ever published. It sounds informed because it's processed millions of examples. It feels generic because it's the average of those examples — not the perspective of a specific person with a specific point of view.
The result is content that could have been written by anyone. Which means it feels like it was written by no one.
What Actually Creates Connection
The content that converts — the article that gets shared, the email that gets replied to, the landing page that closes — has one thing in common: specificity.
Not generic claims, but particular ones. Not "we help businesses grow," but "we helped a Dutch B2B logistics company reduce lead response time from 3 days to 4 hours, which resulted in a 23% increase in conversion rate."
Specificity is what AI can't generate without being given it first. It comes from experience — from actually doing the work, from remembering the client who said something unexpectedly honest, from the pattern you've noticed after working in an industry for years.
The human layer in content isn't about flowery writing or emotional manipulation. It's about the particular — the things only you know because you've actually been there.
The Three Things AI Can't Replace in Good Content
1. Genuine point of view
An opinion is a position that someone could reasonably disagree with. AI avoids these because they're statistically risky — a strong opinion is less likely to be universally agreed with, and AI optimises for agreement.
But opinions are exactly what make content worth reading. "Brand strategy before design" is a point of view. "Rebranding is usually a distraction" is a point of view. "The best websites make a strong impression in 5 seconds or they fail" is a point of view.
People don't share content because it was balanced and comprehensive. They share it because it said something they'd been thinking but hadn't heard articulated before.
2. Real experience
"I worked with a client last year who had the same concern. Here's what happened." You can't prompt AI to say that unless you feed it the story first. And if you're feeding it the story, you're doing the important work already.
Experience-based content is inherently differentiating. No one else has your clients, your projects, your specific observations from the work you've done. When you write from that place, your content is genuinely unique.
3. Earned credibility
AI can state credentials. It can't demonstrate earned expertise through the texture of the writing. The way an experienced practitioner explains a concept — the nuance, the caveats, the things they note from having seen it go wrong — is different from how a text generator explains the same concept.
Readers feel this difference even when they can't name it. It's the difference between content that sounds authoritative and content that is authoritative.
What Good AI-Assisted Content Actually Looks Like
This isn't an argument against using AI. It's an argument against using AI as a replacement for the thinking.
The best approach we've found:
Use AI to create structure. Outlines, section headers, initial drafts that get you past the blank page. Move fast through this stage.
Replace the generic with the specific. Go back through every paragraph and ask: is this something anyone could have written, or does this reflect something we actually know? Replace the generic with the particular.
Add one strong opinion per piece. Find the thing you actually believe — not the safe take, the honest one. Build the piece around it.
Use your own examples. Real clients, real projects, real outcomes with real numbers wherever possible. This is the content no competitor can replicate.
Edit for voice. Does it sound like you — or like a confident but slightly anodyne marketing robot? Ruthlessly remove the latter.
The result should be content that could only have been written by your business. Because it draws on knowledge and experience that only you have.
The Competitive Angle Worth Noting
Here's something worth sitting with: as AI content becomes more prevalent, the differentiation value of genuinely human content increases.
When everything sounds the same, being specific, opinionated, and grounded in real experience becomes a genuine competitive advantage. The market is getting noisier. The signal value of distinctive content is going up, not down.
The businesses that invest in finding and expressing their real point of view will stand out more, not less, as AI content proliferates.