How I Changed My Writing for AI Search (GEO)
Last year I noticed something strange. People I was selling to started showing up with answers from ChatGPT, not Google. Here is how I adapted.
The Bottleneck
When AI search started eating our organic traffic, people panicked. The executives wanted a quick fix, and the SEO team was paralyzed. I had to calm everyone down, scrap our old SEO rulebook, and sit with the writers to completely retrain them on how to write for Generative AI engines instead of Google crawlers.
The Architecture
So I changed how I write. I stopped writing long introductions and started answering the question in the first sentence. I added structured data to every page - the kind that tells AI engines 'this is a tool' or 'this is a how-to guide.' And I stopped publishing things that did not have at least one original data point.
The result: my site started appearing in AI answers for the exact topics I want to be known for. Not because I used some trick. Because I made it easy for AI engines to cite me.
Execution Levers
Shifted our entire content strategy away from keyword stuffing and focused purely on information density.
Ran workshops with the content team to show them exactly how AI language models extract entities and facts.
Stopped writing long fluffy articles and started producing dense structured answers that AI engines immediately started citing.
Target Impact
Organic CTR Decay
Citation Capture Time
Pages with Structured Data
Scale Factors
Stop writing for algorithms and start writing for the question someone just asked.
Start with the answer, back it with your own data.
Tell the search engines exactly what kind of page you wrote with JSON-LD schemas.
Want to Go Deeper?
Every bottleneck is a playbook waiting to happen. If this pattern resonates with a challenge you're facing, I'm always open to a peer conversation.
Start a Conversation