An automated SEO blog for my e-shop.
I built a workflow that prepares and publishes English SEO articles for my own e-shop.
The planned pace is five articles a day, or 35 a week.
By 7 November, 65 articles were live.
The publishing system
1. The article register
A register lists every planned article and its target keyword.
2. Written rules
The agent reads rules for style, article structure, internal links, SEO and what the content may claim.
3. Article generation
An agent prepares each article from the register and rule files.
4. Build and quality checks
A script converts the article into the shop's required format and runs the quality checks.
5. Publication
A GitHub Action publishes the prepared article to the e-shop.
6. Weekly review
A weekly review reads traffic data and changes article priorities.
I manage the system through chat. When a rule needs changing, I tell the agent and update the rule instead of editing articles one by one.
The first three days
| Date | Articles published |
|---|---|
| 5 November | 34 |
| 6 November | 26 |
| 7 November | 5 |
| Total live | 65 |
Another 7 articles were planned and 10 were on hold.
Traffic before the blog
| Metric | Result |
|---|---|
| Active users | 107 |
| Views | 137 |
| Sessions from organic search | 7 |
| Sessions from AI assistants | 3 |
| Purchases | 0 |
This is the baseline for comparing later traffic.
What I learned
The publishing workflow can turn a register of keywords into live articles through written rules, scripts and a GitHub Action.
The next measure is what happens after publication: visits, search traffic and purchases.
What comes next
The weekly review will read traffic results and update the article priorities.