The goal
Checking search visibility, page health and AI crawler activity by hand every week is slow and easy to skip. I wanted a dependable weekly report, and a steady flow of quality articles, without handing my site over to unsupervised AI.
What I built
- A weekly email report: page-by-page health checks, AEO files, crawler access, identity signals, visibility for real client searches, and changes since last week
- Crawler verification: Google, Bing and Apple confirmed by reverse DNS, OpenAI and Perplexity checked against their published IP ranges
- A weekly blog drafting pipeline: the next topic is drafted, built, checked and emailed for review
- Automated quality checks on every draft: title and description length, structured data, answer length, word count and template leftovers
Key decisions
Trust, but verify crawlers
Many requests claim to be AI crawlers. Checking each claim against reverse DNS or the operator's published IP ranges keeps the numbers honest; Anthropic publishes no list, so its visits are only counted on real pages.
Drafts never deploy
Each draft is saved on its own Git branch, and the site only deploys from the main branch, so an unapproved draft can't reach the live site or staging.
Facts come from the source
Prices in drafts are pulled from the live catalogue, and the drafting rules forbid invented clients, results or statistics.