The problem
Crohn's and Colitis Canada (CCC) is Canada's largest volunteer-based IBD charity, supporting 300,000+ Canadians through the MyGUT app and Connect platform. They had years of high-quality patient-education content across 231 web pages, 800+ YouTube videos, and internal research, but no scalable way to turn that knowledge into new articles. Every piece required manual research and writing from a small content team that couldn't keep up.
What we built
We built a Retrieval-Augmented Generation (RAG) content engine. CCC's entire content library was ingested into a Pinecone vector database via Apify and OpenAI embeddings. A Cloudflare Worker hosts the /retrieve API. When CCC's team adds a topic in Airtable, a Make.com workflow pulls relevant context from their own knowledge base, generates a draft tuned to CCC's voice and medical-accuracy standards, and routes it to human review. A scheduled MD5 hash check re-ingests pages whenever the source content changes.
Results
- 231 web pages + 800+ YouTube transcripts indexed as searchable vectors
- Every generated article cites CCC's own approved sources (no generic LLM output)
- Airtable interface — no code, no new tool to learn for CCC's team
- Auto-resync on source page changes via MD5 hash detection