4 weeks of an AI-generated content site: honest numbers and what I've learned
I launched talentedatai.com 4 weeks ago. It's an AI tools review site built entirely with Claude — no coding, no agency, just me and AI.
Here are the honest numbers after one month:
Traffic:
- 206 active users (last 28 days)
- 2.66K Google Search Console impressions
- Average position 9.5 in search results
- Top traffic sources: direct (182), ChatGPT referral (8),
Google organic (6)
Content:
- 21 published articles
- All written by Claude, reviewed and approved by me
- Average 2,500 words per article
- Build time per article: about 20 minutes including research
The unexpected thing:
ChatGPT is sending me traffic. Not a lot — 8 sessions — but it exists. AI search engines are citing my AI-generated content about AI tools. The recursion is not lost on me.
What's actually working:
The Gemma 4 Android guide is getting organic search traffic for "how to run Gemma 4 on Android phone 2026." It's a specific, practical guide that answers a specific question. That's the pattern — specific beats general every time.
What's not working yet:
CTR is 0.2%. I'm getting impressions but people aren't clicking. The meta titles and descriptions needed work — fixed this week. Will report back in 30 days.
The honest question I'm sitting with:
The site is AI-generated content reviewed by a human. Is that enough? Google says it doesn't care about how content is produced, only whether it's helpful. So far the data says it's being found and read. Time will tell if it builds authority.
What I'm doing differently in month two:
Writing fewer, more specific articles rather than broad coverage
Focusing on comparison queries where I can see search demand in Search Console before writing
Building the newsletter as the real long-term asset
Will post a 60-day update. Happy to answer questions about the setup, the numbers, or the experiment. I would really appreciate if you visit the site and sign up for the AI guide and ongoing newsletters!
Site: talentedatai.com
Really interesting seeing the actual numbers after 4 weeks, especially what worked vs what didn’t. Curious, did you mainly use Claude for the whole workflow, or did other AI tools play a part too?
Firstly: most people on week 4 are still ‘perfecting their idea’. Congrats on taking action fast (and it clearly paid off).
You mentioned that you are planning on utilizing a newsletter and articles.
Just curious: do you think it would help if you could have someone take the consistent writing part off your shoulders (while still maintaining your voice in the posts that did best and persuasive quality) so you can focus on the features.
If that sounds like something you would be interested in (even a little bit), I can show you some samples.
Either way, CONGRATS ON THE PROGRESS.
The ChatGPT referral traffic part is the most interesting number in your whole post. 8 sessions from ChatGPT in 4 weeks — that's tiny, but it's a signal most people completely ignore.
Here's what I'd want to know: when someone asks ChatGPT "best AI tools for X," does it actually mention your site? Or is the traffic coming from ChatGPT linking to a specific article? That distinction matters a lot for your month 2 strategy.
Because if ChatGPT is citing your specific how-to guides (like the Gemma 4 one), that's a completely different growth lever than Google SEO. Google rewards backlinks and domain authority. AI chatbots seem to reward clear, structured, specific answers — which is exactly what you're already producing.
The recursion you mentioned is real but it goes deeper: if you optimize your content to be cited by AI, you might get more AI referral traffic than Google organic. At 6 Google organic vs 8 ChatGPT referrals, AI is already winning for you.
Curious — have you checked whether Perplexity or Gemini also cite your articles? Would be interesting to compare across AI engines.
the honest question you left at the end is the right frame: 'AI-reviewed - is that enough?' I'd say yes if you'd stake your name on it. Google's right to not care about the draft source.
The ChatGPT referral traffic is honestly the most interesting data point here. 8 sessions is tiny but the fact that it exists at all is a signal most people would miss. AI engines are citing your content, which is basically a brand new distribution channel that works on completely different rules than Google.
For the CTR thing - 0.2% at an average position of 9.5 isn't actually that bad for month one. Almost nobody clicks past position 3. The fact that you're indexed and ranking already is the hard part. As you narrow your focus to specific queries like the Gemma guide, you'll naturally move up and CTR follows.
Curious to see if the ChatGPT traffic grows as you add more specific guides.
Early signal looks promising.
The biggest insight here: distribution is starting to shift—AI tools sending traffic to AI-generated content is a glimpse of what’s coming.
Also, classic lesson playing out again:
specific, problem-solving content > broad, generic posts.
If you can fix CTR and build an email list alongside this, you’re not just building traffic—you’re building an asset.
Super interesting breakdown — especially the “specific > general” insight, that’s been true everywhere lately.
Also agree on CTR — small tweaks there can change everything. Curious to see your 60-day update.
Also, if you’re experimenting with ideas — you have an idea, $19 puts it in real competition. Winner gets a Tokyo trip (flights + hotel, $500 min guaranteed).
Round just opened: tokyolore.com 🚀
This is a solid start, most people never even get indexed in month one.
The shift you’re making toward specific queries and distribution is exactly what turns this from content into an actual asset.
GOood
I like your app Sunsur. There are a lot of big AI portals discussing many AI topics, you are adding value to that stack.
If you are interested write an article about Travel Apps.
I developed one lupath.ai about travel inspiration. I would like to read about Travel Apps status in 2026.
Cheers!