Home › A weekly AI SEO page pipeline that stages pages for approval
A weekly AI SEO page pipeline that stages pages for approval
Affiliate disclosure: Mulkern AI Systems is an independent affiliate partner of Easy (easy.one), a product of Smart Apps Limited. If you buy Easy through our links, Mike Mulkern earns a commission, and the coupon applied at checkout (MICHAEL) is his. Everything on this page describes how we run our own businesses on Easy. It is our experience, not a statement by Easy.
Every Wednesday at 6:00 Central, a scheduled agent on our Easy installation generates a batch of SEO pages, scores each one, runs them through a stack of automated gates, and stages the result. It does not publish. Since 30 September 2026, applying a batch needs Mike's approval. This page covers how the SEO page pipeline runs and what we learned about checking AI-written pages.
What runs each week
The pipeline has three scheduled lanes: pages for TrafficForge on Wednesdays, pages for Oil Rigs Now on Fridays, and a weekly blog post on Wednesday afternoons. A separate SEO agent runs a weekly audit on a six-week rotation, one property at a time, so each site gets a deep look rather than a skim.
Each page is scored on sixteen criteria (title, description, headings, length, structured data, links, and so on) before the gates run.
Staging is the successful outcome
The weekly brief says it plainly: generate, score, gate, and stage, then stop before the commit. Staging is success. When the batch is approved, applying it publishes the exact bytes that were reviewed and generates nothing new.
That last rule matters. If approval triggered a fresh generation, the published pages would be different from the ones a person checked.
A documentation fix is not a gate
The most useful lesson came from a claim the product did not support. It sat in a page template, and it was corrected in the documentation, but the generator kept producing it. A correction written in a note does nothing to the next run.
So each recurring mistake became an automated gate that reads the finished pages and fails the run: one for unsupported product claims, one for a canonical tag on the wrong host, one for pages no other page links to, and one that reads the generator's own template text, because a claim hardcoded in a template looks like intended design on every page.
A gate that scanned nothing must never look like a pass, so every gate has a separate result for "nothing scanned."
Running a pipeline like this on Easy
On Easy, the schedule is a scheduled task, the agent is briefed in plain English, and the scripts the agent writes live in its workspace. The pages on this site went through the same gates.
If you build one, keep a record of the checks each run actually performed, not only the ones it was meant to perform. We assert that the list of gates invoked equals the list expected, so a dropped check fails loudly.
In the Easy documentation, see Make it repeat and Workspaces.
Questions
Does the pipeline publish without a person?
No. It stages pages, and Mike approves a batch before it is applied. Approved content changes follow a written rule with an undo recorded before anything ships.
Which scoring system do the pages use?
QualityForge Lite, the sixteen-criterion scorer from TrafficForge, plus several site-level gates we added for problems a single-page score cannot see.
Can Easy run scripts as part of a pipeline?
Our agents write and run scripts in their own workspace. Check the Easy docs for what your plan and installation allow.
Is this page sponsored by Easy?
No. It is an affiliate page written by Mulkern AI Systems. Mike earns a commission on purchases made through our links.
More from how we run on Easy
Easy is the platform both of our companies run on. Our main page explains what it is, compares the three plans, and applies our coupon at checkout.
See the plans on our Easy page