Intelligent Content Repurposing With Approval Workflow Case Study
An AI-powered content repurposing agent for content teams. It transforms long-form content into platform-optimized posts for LinkedIn, Threads, and Instagram with matching AI-generated visuals, routes everything through Slack for one-click human approval, and publishes across all platforms simultaneously via Buffer.
Service
Automation
Industry
SaaS & Tech
Category
AI Automation
Built with: n8n, Llama, Slack, Buffer.
Overview
The Problem
Content creators struggle to adapt single pieces of content into multiple platform-specific formats efficiently
Generating platform-specific copy (LinkedIn, Twitter, Instagram) requires different tones, lengths, and messaging strategies
Creating matching visuals for social posts is time-consuming and often outsourced to designers
Without a centralized approval workflow, inconsistent or off-brand content gets published
No audit trail of content decisions, making compliance and performance tracking difficult
The Solution
AI instantly transforms long-form content into three distinct platform-optimized social posts
Automated AI-generated graphics ensure visual consistency across all platforms without designer overhead
Slack-based approval workflow keeps stakeholders in the loop before publishing
Once approved, Buffer schedules posts across LinkedIn, Threads, and Instagram simultaneously
Every decision is logged in Slack for full compliance and content performance tracking
Solution Workflow
Submit
Generate
Visual
Review
Approve
Publish
Full Technical Breakdown
1. Flexible Content Input Interface
Accepts content via an n8n web form in two formats, direct paste or file upload (PDF, DOCX, TXT), and detects which method was used to route accordingly.
2. Intelligent Input Normalization
A JavaScript code node analyzes the submission, detects whether a file was uploaded or text was pasted, and normalizes the input structure for consistent downstream processing.
3. Conditional File Extraction
An If node routes file submissions to text extraction from PDF/DOCX/TXT, or passes pasted content directly through, both paths merge into a standardized Final Content node.
4. AI-Powered Multi-Format Content Generation
Llama 3.2 (3B) generates a LinkedIn post (professional, 150-300 words), a Twitter/Threads post (max 150 characters), and an Instagram caption (conversational, hashtags), prompted to return JSON only for reliable parsing.
5. Sophisticated AI Output Parsing
A custom JavaScript parser strips markdown code fences, extracts content via regex, converts escaped characters back to readable text, and falls back to default messages if extraction fails.
6. Automated AI-Generated Visual Asset
Pollinations AI generates a 1024x1024 minimalist professional graphic using the first 100 characters of the LinkedIn post as the prompt, stored for Slack preview.
7. Slack-Based Approval Workflow
n8n builds a rich Slack message with section blocks per platform, the AI-generated image, and Approve/Reject action buttons, making approval a one-click decision.
8. Action Capture & Routing
Slack sends a webhook callback on button click; n8n parses the action type and routes to Buffer scheduling on approval or a revise-and-resubmit notification on rejection.
9. Multi-Platform Buffer Scheduling
Three parallel Buffer nodes schedule the LinkedIn, Threads, and Instagram posts with the AI-generated graphic, each optimized for its platform and audience.
10. Confirmation & Audit Logging
Slack receives a scheduling confirmation, and timestamp plus user information are logged with the full message history preserved for compliance.
Results
Key Features
Dual input methods: Users can paste content directly or upload PDF, DOCX, or TXT files, no need for multiple submission methods or technical setup.
Automated format adaptation: Generates LinkedIn (professional), Threads (concise), and Instagram (casual, hashtags) versions automatically, each optimized for its platform.
AI-generated visual assets: Eliminates the need for manual graphic design by automatically creating on-brand, professional social media graphics.
Slack-native approval workflow: Rich, interactive Slack messages with embedded content preview and one-click approve/reject buttons, approval happens where teams already communicate.
Multi-platform scheduling: Upon approval, Buffer automatically schedules posts to LinkedIn, Threads, and Instagram simultaneously with consistent messaging.
Resilient AI output parsing: Custom JavaScript parsing handles imperfect AI responses, including escaped characters and markdown fences, ensuring reliability across model outputs.
Before
Manual rewriting of content for each platform takes hours, inconsistent messaging and tone across channels, graphic creation outsourced or manually designed, no centralized approval mechanism, posts scheduled manually, zero audit trail of content decisions.
After
All three social formats generated in minutes via AI, consistent messaging guaranteed by AI and human approval, professional graphics auto-generated for every post, mandatory Slack approval before publishing, all platforms published simultaneously via Buffer, complete audit trail of all decisions.
What This Proves
This system demonstrates that AI can handle content transformation, format adaptation, and visual generation while maintaining quality through human-in-the-loop approval, freeing content teams to focus on strategy and creative direction rather than repetitive formatting and scheduling tasks.
Tech Stack
Content Intake
n8n Form Trigger — Web form for content paste and file upload
File Extraction — Automatic text extraction from PDFs, DOCX, TXT
AI & Content Layer
Llama 3.2 (3B) — Local LLM for multi-format content generation
Pollinations AI — Automated AI image generation for social graphics
Slack — Approval workflow, notifications, and audit logging
Scalability & Flexibility
Content Source Flexibility: Currently accepts manual paste and file upload; can extend to integrate with WordPress, Medium, Substack, Notion, or RSS feeds to auto-trigger repurposing.
Platform Expansion: Currently targets LinkedIn, Threads, and Instagram via Buffer. Easily add TikTok, YouTube Shorts, Bluesky, Mastodon, or custom CMS integrations by adding new Buffer or native API nodes.
Multi-Language Support: Modify AI prompts to generate content in different languages, enabling global distribution of the same source content.
Custom AI Prompts by Content Type: Create template prompts for different content types (blog post, whitepaper, case study, announcement, product launch) selectable at submission time.
Advanced Approval Chains: Add multiple sequential approval steps (brand manager → compliance → legal) using Slack's workflow builder or additional approval channels in n8n.
A/B Testing Variants: Generate multiple AI versions and surface them in Slack, allowing approvers to choose the best variant before publishing.
Engagement Analytics Feedback Loop: Integrate Buffer's post-publish analytics (likes, comments, shares) back into n8n to identify top-performing formats and optimize future content.
Content Performance Dashboard: Automatically log all approved content and AI variations into Airtable or Google Sheets for historical analysis and competitive benchmarking.