𝗪𝗵𝘆 𝗜 𝗮𝗺 𝗕𝗲𝘁𝘁𝗶𝗻𝗴 𝗼𝗻 𝗗𝗶𝗿𝗲𝗰𝘁𝗼𝗿𝗶𝗲𝘀 𝗼𝘃𝗲𝗿 𝗚𝗼𝗼𝗴𝗹𝗲 𝗔𝗜 𝗢𝘃𝗲𝗿𝘃𝗶𝗲𝘄𝘀
Google AI Overviews answer many search queries without a click. People ask for the best AI tools, and Google gives a list right on the search page.
This looks like a death sentence for my directory sites: Top AI Tools, Find Games Like, and Open Alternative To.
I launched these sites on April 23, 2026. My total cost is $25 per month. I have a specific bet. By October 2026, at least one site will get 200 organic clicks per month for two months straight. If this fails, I will publish my data and admit I was wrong.
Why do I think directories can still win?
AI Overviews are good at summarizing text. They are weak at three specific things:
• Attribute Filtering: AI gives prose. My sites use structured data. You can filter for tools that work offline or have a mobile app. You cannot easily do that with a paragraph of text.
• Negative Space: AI usually stays positive. My sites include "avoid if" sections. This tells you who should skip a tool.
• Freshness: AI relies on web mentions, which can be old. My system checks GitHub activity weekly. I know if a tool is actually being maintained.
I am also targeting the "downstream" search.
A user might see a list in an AI Overview. Then, they search for a specific comparison like "Appflowy vs Anytype." They want a verdict, not a summary. My sites provide structured comparison data for these high-intent users.
Here is how the two compare:
• Discovery: AI is strong. Directories are weak. • Comparison: AI is medium. Directories are strong. • Filtering: AI is low. Directories are high. • Freshness: AI is inconsistent. Directories are high.
Running this experiment costs very little. I use Vercel, Turso, and Claude Haiku. This low cost lets me stay objective. I do not need to force a positive result to stay alive.
I will watch for three failure signs:
- High impressions but zero clicks on comparison pages.
- Google AdSense rejecting the sites even after I improve the content.
- People stop using Google search for comparisons and move entirely to ChatGPT.
I will share the results with raw screenshots in October 2026.
Optional learning community: https://t.me/GyaanSetuAi