How We Make Our Content

SEO Lab posts are drafted with Claude, Anthropic’s AI model, then checked and published by Rob Boirun. Every number in a post comes from an export, a file or a page we can point to.

What does Claude do, and what does Rob do?

1. Rob picks the project and the question the post has to answer.
2. Claude reads the evidence: the repository, its commit history, the Search Console or GA4 export, and the official documentation. It drafts the post and a separate list of where every number came from.
3. Rob reviews the draft against that list and asks for changes. Nothing publishes without his approval.

Where do the numbers come from?

• Search Console and GA4 exports from sites Rob runs, with the date range stated in the post.
• Files and git history in his own repositories, including commit dates.
• Crawls and curl checks run on a date the post states.
• Official documentation from Google Search Central, Next.js and Bing, fetched while the post was written. The Sources section links each page.

When the data does not answer a question, the post says so and names where to confirm it. We do not estimate into a gap.

Are the images AI-generated?

No. Featured images are screenshots of the real sites, taken with a headless browser. Screenshots of Search Console, GA4, terminals or code are real. If we ever publish an AI-generated image, the file will carry the IPTC DigitalSourceType value trainedAlgorithmicMedia.

What about client sites?

Some projects are client work. A post names a client or shows a client’s data only with the client’s permission, and never includes anyone’s personal details.

How are corrections handled?

When a fact in a post is wrong, we fix it in the post and add a dated note saying what changed. To report an error, leave a comment on the post.

Sources

• Google Search Central, Google Search’s guidance on generative AI content on your website.