AI-written ads tend to get more clicks. Human-written ads tend to close more sales. Both of those things are true at once, and the interesting part is why.
What the pattern actually looks like
AI copy typically pulls around 12% higher CTR, but human copy often drives roughly 60% more sales clicks across the same campaigns. AI hooks people. Humans close them. The gap widens further on high-ticket offers, where AI-only copy can underperform by 8–18%, while simpler, lower-stakes offers sometimes do just fine running on AI alone.
Why AI wins the top of the funnel
AI can generate large volumes of variations fast, which means it finds novel hooks and unusual angles quicker than any human writer working alone. It’s genuinely strong at pattern recognition across data. What it lacks is a feel for the emotional arc of a buying journey — it can write a headline that stops the scroll, but that same output can misalign with what actually needs to happen right before someone decides to buy.
Why humans win the close
Human writers pick up on nuance and trust in a way that still outperforms AI, especially on high-ticket offers where the sale depends on handling specific objections and building real credibility. That’s subtle buyer psychology that tends to get lost in AI output precisely because it requires reading between the lines rather than optimizing for an obvious signal.
The workflow that actually beats either one alone
Rather than picking a side, the workflow worth building uses both deliberately: AI drafts a large batch of variations fast, humans pick the strongest of those, and then a human writer rewrites and sharpens the chosen hooks before they go live. That combination gets you AI’s speed at generating options and human judgment on which ones are actually worth running — and it tends to lift both CTR and conversion together rather than trading one for the other.
How to set this up in practice
Brief the AI with real constraints and your actual brand voice, then generate ten to fifty hooks per concept rather than settling for the first few. Filter with simple metrics plus human review to get down to three to five strong candidates. From there, a human writer builds out the final ad copy and the landing page experience together — checking that the headline and the landing page actually say the same thing is worth doing every time, since mismatches there quietly kill conversion.
Worth reading further
Harvard Business Review has good research on where automation and human judgment each add value, Google Ads Support covers practical setup guidance, and OpenAI is a useful source on where current AI capability actually sits if you want to calibrate expectations.
Where to start if you want to test this
Start small — one campaign, split traffic, and measure both CTR and what happens downstream. Document exactly what the human rewriter changes from the AI draft; that record becomes the playbook for scaling the process across other campaigns.
Related reading on this blog:
- Scale Ads by Incremental Attribution
- 7 Skills That Replaced Our Ads Workflow
- Facebook Ads Limits That Kill Performance
- High-Converting Google Ad Anatomy
- Free Claude Skills for Paid Ads
The takeaway isn’t to pick a side. It’s to let AI generate the hooks at scale and let humans handle the close — that combination is what actually drives clicks and sales together.