Case Studies
These aren't ChatGPT prompts. They're production systems with APIs, databases, quality controls, and measurable ROI.
Zimbabwe's largest property portal needed to turn 4,000+ high-intent keywords into quality content — without the slop, without the six-figure freelance budget, and without losing their unique competitive edge (a private database of 7,000+ listings).
I built a 6-step AI pipeline: keyword selection via GSC/ahrefs, competitor scraping, private DB integration, EEAT assembly with Google Maps data, Claude generation + automated QA, and a full CMS with performance tracking.
A career platform wanted to offer AI-powered resume rewrites — but the manual process cost $240 per rewrite and took days. They needed something that could scale, maintain quality, and convert free users to paid customers.
I built a multi-layered AI system: resume parsing, LinkedIn scraping, live job posting analysis, ATS keyword extraction, dynamic writing guideline selection, multi-stage QA, and Google Docs export with embedded edit suggestions — all in a 58-table relational database architecture.
Real estate agents at Pam Golding Properties were spending 45+ minutes per listing writing descriptions — a tedious, repetitive task that produced inconsistent quality and took agents away from actually selling property.
I built a Chrome extension that injects directly into their existing Property Book system. One click: the extension scrapes the listing, analyzes property images with Vision AI, selects writing guidelines via RAG, and returns a brand-consistent description in under 20 seconds.