Content Inputs — The Content Factory

Purpose

Raw content materials that feed the AI Production Agent as source data for content generation. Distinct from knowledge-base/ (which holds brand voice, positioning, and stable brand intelligence) — this folder holds the SPECIFIC MATERIAL a given piece of content is ABOUT.

Typical inputs: product information, blog posts, ingredient breakdowns, customer testimonials, research citations, existing social posts, raw notes, photo/video asset lists.

Organization

One subfolder per product (or per topic, if topic-led rather than product-led):

content-inputs/
├── README.md                (this file)
├── [product-name-kebab]/
│   ├── product-info.md       primary product description, positioning, claims
│   ├── ingredients.md        ingredient list, sourcing, mechanism
│   ├── blogs/                one .md per blog post
│   ├── testimonials.md       customer quotes (anonymized if needed)
│   ├── existing-social/      screenshots + captions of posts already out
│   └── assets-manifest.md    list of available photos/videos with paths

File Conventions

• Markdown preferred for text content (converts cleanly from Word/docs via pandoc).
• Use kebab-case filenames.
• When pasting blog content, include title, author, publish date, URL (if live), and body.
• For product-info: start with one-paragraph product summary, then positioning, then claims (only claims present in existing ZC docs — don't invent).
• For testimonials: quote + customer first name + product used + date (or "ongoing"). Anonymize if needed.
• For assets: list paths, don't copy large binary files into this folder.

How the AI Uses These

1. The AI Production Agent reads knowledge-base/ to understand brand voice, positioning, personas, pillars (stable brand intelligence).
2. It reads content-inputs/[product]/ to understand the SPECIFIC material this piece is about.
3. It produces outputs per owner-track (Alice educational, Gloria visual / brand world, Daiana newsletter and affiliate, Lathifa podcasts and lives, Zuman content direction, Gisela pitches and audiences) that fuse brand voice with product material.
4. Outputs drop into synthesis/outputs/[product]/[owner-track]/ (created as needed in Phase 3).

Pilot Status (2026-04-18)

First pilot product is being selected by Can. When files arrive, place them in content-inputs/[product-name]/. The pilot will exercise the 1-product → 4-owner-track generation pipeline.