Back in 2022, Simplebo sold its customers a five-page content extension, and the Customer Success team wrote every page by hand. The Board wanted more of it sold, and hand-writing would never keep up. Early in 2023, the tools that could close that gap had just arrived, which is the terrain behind Content at scale.
TL;DR
I built a Content Factory inside Simplebo’s marketing team to help produce the five-page extensions at industrial volume, with a human review that made the output publishable. The Board pulled the design back from full automation to AI as a copy assistant, and that step-back is part of the method. What I proposed and what shipped were different numbers, and I state both.
The situation I found at Simplebo
The Board discussed, at director level, how hard it had become to produce the web pages customers bought as an extension. The Customer Success team owned that work, and writing for the web was not their specialty. Meanwhile, the ChatGPT boom of early 2023 put a wave of SaaS web-content tools in front of every marketing team; one of mine was testing them hard to write Simplebo’s blog articles.
The pages were not the problem. The people writing them had another job.
The diagnosis: production was the constraint
I read the problem as production, not demand: the pages were written slowly, by hand, by a team whose specialty was customer relationships, and web production belonged with people trained to do it, the marketing team I led.
The Board read it differently. For them the constraint was demand, and the risk was brand trust: if a customer thought some pages were AI-assisted and lost confidence, the extension’s value fell.
Two readings, and the decision had to answer both.
What I proposed, and what the Board approved
I proposed taking production off the Customer Success team and running it as a Content Factory inside marketing: templates per profession, a customer onboarding questionnaire, a quality process, and a human review at the end. What carried the decision was the arithmetic of the change .
The Board approved the move, not the full design. Back in 2023, using too much AI-assistance to create pages could be an issue for some customers. So, the Board asked for a less powerful approach: original content for each customer instead of templates, with AI as a content writer assistant. The full-automation design never shipped, though production still moved onto the Content Factory.
What the Content Factory actually was
What went into production was narrower than the design. I led the marketing team that built and ran it, with a champion named inside Customer Success. The first version was tested on five representative personas, one per profession, so the test carried Simplebo’s diversity.
The factory held four things:
- Templates per profession. Each had its own starting structure, and the library grew by adding use cases inside each one.
- A customer onboarding questionnaire. A Customer Success teammate runs it during onboarding, and the answers personalize the templates.
- A collision-avoidance routine. Two sites in the same profession and service area never share the same pages.
- A quality process. A prompt library, checklists, a dedicated copy-writer, and a supervising chief content editor, with the Customer Success champion validating the templates.
What broke in production
Three things went wrong, and none was the one the Board feared. The human operator had to stay vigilant over what the model produced because in 2023 the context window was thin. The prompt library was hard work to maintain, and when it fell short, the failure showed up as rework: test and learn, and pages the operator had to rebuild.
The checklist was not bureaucracy: a page reached a customer’s website only if Customer Success took it there.
The outcome, with its status
Here are the honest numbers. What I sold the Board was a proposed capability, not a delivery figure: a rough 80 websites and 800 pages of content per month, the maximum production capability of the one person dedicated to it. It was never reached. What actually shipped allowed selling 20 to 40 extension packages per month.
My Marketing Director mission ran from June 2022 to June 2023, and I have no visibility into what Simplebo did with the Content Factory after that, so I claim no later outcome. The human review at the end held, and it is the part I would repeat: industrial-scale AI content and brand trust are compatible when a person signs off on what ships .
The operator behind this method is on the About page. The Simplebo engagement is also a career mission, told on my dedicated website: operational relevance in the age of AI, evidenced twice.
If this is the shape of your problem, let’s talk.