Automated SEO Content Pipeline

The challenge

OASIS212 is a kitesurf and wellness resort in Dakhla, Morocco. The brand needed editorial content that could establish experience, expertise, authority, and trust in a category where local depth, niche expertise, and SEO discipline rarely come from the same person. Dakhla, kitesurfing, and wellness together is a narrow enough combination that freelance options meant choosing between a generalist with polish but no local depth, or a specialist with real expertise but content that wasn’t built for search.

The approach

Rather than treating this as a writing problem, I built a system that could consistently turn local knowledge, brand voice, and SEO logic into useful content. The local knowledge wasn’t generated by the model. I built that layer by gathering insight from people who actually know Dakhla: local kitesurf instructors, wellness practitioners, guides, and people from the region. Their knowledge, alongside the brand’s voice and editorial requirements, became part of the system’s foundation. From there, AI handles the repetitive parts of the workflow — research, structuring, drafting, and formatting — while human judgment remains central to the process. I deliberately kept publishing out of the automation loop, so every piece can still be reviewed, challenged, and refined before it goes live. The result isn’t a system designed to replace writers or remove editorial judgment. It’s a system designed to give human expertise more leverage: capturing valuable local knowledge once, translating it into a repeatable content framework, and using AI to help apply that thinking consistently week after week.

The execution

1. 01The automated pipeline generates 3 topics across seven themes covering the brand’s core categories
2. 02Checks topics against a maintained history so the system doesn’t repeat angles cycle over cycle, with an automated feedback-loop pulling directly from the workflow’s own output history
3. 03Runs competitive research per topic: pulls the top-ranking pages for the target keyword and analyzes their structure, so every brief is built against what’s actually winning that search
4. 04Sends each brief, along with a detailed brand and voice rulebook covering tone, positioning, and formatting, to the writing model in one structured prompt, so every article matches brand voice without manual rewriting
5. 05Returns a full SEO outline, the complete articles, meta titles, descriptions, and slugs in a single output
6. 06Routes every draft to Google Docs for editorial sign-off before anything reaches the CMS. No auto-publish