I use AI (Claude) a lot for work on my website. This note covers the 3 rules I set so the AI works fast but nothing wrong goes public on the site.
Key takeaways
- New AI-made pages start as drafts and go live only after a human check.
- Never publish fake numbers, reviews or rates; leave the spot empty or say “message me on WhatsApp”.
- First-hand experience sections come from me, not from the AI.
1. Draft first, then publish
New pages start as drafts. They go public only after I have checked them and details like contact info are correct.
2. No fake numbers, reviews or rates
Where a rate or an audience number is not decided, the page says “message me on WhatsApp” or leaves the spot empty. A fake number breaks trust and can also hurt you with AdSense.
What Google says
Google’s guidance focuses on whether content is helpful and reliable, not on whether AI was used. First-hand experience is one of the clearest signs of helpful content. Source: Google Search Central: Creating helpful, reliable, people-first content.
3. Real experience comes from me
AI can build the structure and the first draft, but sections like “what went wrong and what I learned” only contain what I saw myself. That is the part Google and readers notice.
Frequently Asked Questions
Is it OK to use AI to write website pages?
Yes, as a helper. Use it for structure and first drafts, check every detail yourself, and keep the page as a draft until it is correct.
What should never be AI-made on a website?
Numbers, reviews, rates and personal experience. If a number is not decided yet, leave it out or ask people to message you.
References
Official documentation and research used for this article:
- Google Search Central Blog: Google Search’s guidance about AI-generated content – AI content is fine when it is helpful and accurate.
- Google Search Central: Creating helpful, reliable, people-first content – first-hand experience and trust.
- Google Search Central: Spam policies – what Google treats as spam, including scaled low-value content.