Insights
Notes from running AI automation in production.
What I learn building and operating content pipelines and websites every day: what works, what breaks and what it costs.
The period-length trap in Search Console reports
How comparing two unequal date ranges made a growing site look like it was shrinking, and the simple rule that prevents it.
In AI content, the quality gate is the product
Writing is the easy part. The real work is catching what the AI gets wrong before it goes live, including the mistakes the checker makes itself.
What an AI content pipeline really costs to run
The real per-post cost of publishing with AI, where the money actually goes, and the changes that halved my keyword-data bill.