Built an automated, OpenAI-driven content engine — SEMrush gap analysis, Zapier automation, and human review — that lifted impressions 152% and expanded top-ranking keywords by hundreds over six months.
The project aimed to revamp our organic content and SEO strategy by building an automated content engine on top of OpenAI. Rather than relying on a one-off content calendar, I designed a repeatable pipeline that finds real keyword gaps, drafts on-brand content to fill them, and routes everything through human review before anything publishes — as part of a broader organic content growth strategy.
Workflow
Research
SEMrush surfaces keyword gaps against competitors and our own content library. Human review filters the list for topical relevance and likeness before anything moves downstream.
Automation
Gap-analysis data feeds into Zapier, which kicks off an automation that maps target keywords and density against the existing sitemap — pinpointing exactly where content gaps exist.
Generation
A prompt fires against OpenAI to generate unique, keyword-rich content, guided by a voice.md style reference so the output reads in a consistent, human voice rather than generic AI copy.
Drafting
Finished drafts land directly in WordPress as pending posts — ready and waiting for editorial review inside the CMS.
QA
Every draft passes through a human review gate: publish, revise, or reject. Editorial judgment is the final check before anything goes live.
Loop
Published content re-enters the pipeline on a cadence, re-running the keyword gap analysis to catch new gaps and refresh aging pages automatically.
Keyword Gap Analysis — Used SEMrush to identify keyword gaps against competitors and our own content library, paired with human review to judge topical relevance and content likeness before anything moved downstream.
Automated Ingest & Mapping — Fed gap-analysis data into Zapier, which triggered an automation mapping target keywords and density to our existing sitemap — pinpointing exactly where content gaps existed and what needed to be written.
AI Draft Generation — Automatically initiated prompts against OpenAI to
generate unique, keyword-rich drafts. A voice.md style reference guided
every generation so the output read in a consistent, human voice rather than
generic AI copy.
Draft Delivery & Human Review — Generated drafts landed directly in WordPress as pending posts, ready for editorial review — keeping a human reviewer in the loop on every single piece before publication.
Review & Approval — Every draft went through a review and approval gate: publish, revise, or reject, with editorial judgment as the final check before anything went live.
Continuous Refresh Loop — Published content re-entered the pipeline on a cadence, re-running the keyword gap analysis to catch new gaps and refresh aging pages — turning the engine into a self-sustaining system rather than a one-time push.
The impact of the AI content engine is reflected in the substantial improvements in our key performance metrics over a six-month period:
These metrics highlight the effectiveness of pairing OpenAI with a disciplined, human-reviewed workflow — boosting organic search performance and driving meaningful engagement with our target audiences. Through this project, we transformed our digital presence and built a durable, repeatable system for continued growth.