Chapter 12: Creative Workflows
Creative work has a dirty secret: most creative blocks are logistical, not creative. The actual creative work—making it good—is human. But gathering references, organizing research, reformatting content, distributing finished work, tracking engagement—those are all automatable, and automating them gives you more time and headspace for the creative parts. The content research-to-draft pipeline is the core pattern: web search across multiple queries, fetch and read top results, save research notes, write draft from research, send WhatsApp when ready. The human touch-point comes after the first draft ("I'll specify" is explicit in the workflow). Content calendar automation on Monday mornings reads the week's planned content, creates research briefs for undrafted pieces with search queries, target length, and audience.
Image and media organization handles the file chaos that creative work generates. Photo organization pipelines use the vision model to analyze incoming images, generate descriptive filenames, categorize (portrait/landscape/product/screenshot/document), move to organized subdirectories, and update an index. Screenshot cataloger cron jobs find today's screenshots in Downloads, analyze with vision, categorize and rename, move to organized archive, and delete junk. Asset versioning copies existing files to archive with timestamp before saving new versions, updating a CHANGELOG. Social media automation drafts threads, LinkedIn posts, and Instagram captions from content ideas, with engagement tracking via platform APIs and content recycling (finding old high-performing posts and generating fresh platform-specific versions). The writing assistance pipeline breaks into explicit phases: research (search + summarize), outline (3 options to choose from), then draft using chosen outline and writing style reference—human check-in between phases.
Podcast and video workflows automate the mechanical parts: show notes generation from transcripts (summary, key takeaways, timestamps, guest bio, resources, SEO title variants), YouTube publish packages (description, tags, chapters, 5 title options, Twitter thread, LinkedIn post), and transcript processing that auto-detects content type and applies the appropriate template. The content machine case study is a complete content production system: Monday morning automation reads ideas.md and calendar.md, creates draft briefs, moves approved items to active projects, and sends weekly plan; daily drafting runs research phase and generates first drafts moving them to review; publishing pipeline triggers on file movement from review to approved—final formatting, word count, SEO analysis, social post generation, scheduling. Human job: review and approve. Everything else automated.
Key Patterns
- **Research → outline → draft with human check-in:** Automate gather and organize; human does "make it good"
- **Content machine with directory lifecycle:** ideas.md → queue/ → drafts/ → review/ → approved/ → published/
- **Vision-powered file organization:** Analyze images, rename descriptively, categorize, move to organized structure
- **Monday content prep automation:** Read calendar, create research briefs, send weekly plan via WhatsApp
- **Content recycling:** Find old top-performing posts, regenerate for each social platform with fresh angles
Related Concepts
- [[kelly-handbook-ch11-business-automation]] for the document processing pipelines that feed creative work
- [[kelly-handbook-ch6-cron]] for the Monday morning content prep scheduling
- [[kelly-handbook-ch5-communication-automation]] for WhatsApp-based draft review and approval
- [[kelly-tweets-openclaw]] for social media and content automation patterns from the Kelly corpus