WZ
William ZainEnterprise IT x Digital Creative
← All case studies

Case Study: Autonomous Multi-Channel AI Content & Media Distribution Engine

The Business Challenge (The Friction)

Creating, formatting, rendering, and distributing high-value technical tutorials across multiple social media platforms was a major operational bottleneck. Producing educational AI breakdowns across Facebook, LinkedIn, YouTube, and internal archives required manual topic selection, copywriting, AI prompt engineering, media rendering, and manual uploading.

This fragmented process caused substantial friction:

  • Manual Overhead: Consumed 8 to 10 hours of manual administrative and operational effort per week.
  • Channel Formatting Inconsistencies: Copy pasted between tools frequently broke platform formatting (e.g., unparsed markdown asterisks and missing line breaks).
  • Delayed Engagement Execution: Manually posting links in comments to bypass algorithmic penalties on main posts caused inconsistent distribution timing.
  • Operational Disconnection: Educational assets were rarely archived into a centralized knowledge base in real time.

The Solution Architecture (The Build)

Screenshot 2026-08-29 150505.png

Screenshot 2026-08-29 150505.png
Screenshot 2026-08-29 150505.png

Built on Make.com, the workflow integrates Google Gemini API, Facebook Pages API, YouTube Data API, LinkedIn API, and Notion into a unified, headless publishing pipeline.

Trigger & Dynamic Ingestion: Executes on a scheduled cadence (Tuesday/Friday at 1:05 PM). It dispatches an execution start email notification, randomly selects an advanced technical topic from a rotating array (e.g., motion brush mechanics, multi-camera speed ramping, character consistency), and injects global call-to-action (CTA) destination links.

  • AI Analysis & Structured Reasoning: Feeds the dynamic context into Gemini 3.5 Flash using a strict system instruction and a mandatory JSON output schema. The model generates four discrete outputs: a click-worthy headline (Output_1_Headline), optimized post copy with native CTA placement and 5 hashtags (Output_2_Facebook_Post_Copy), a production-ready rendering prompt (Output_3_Image_or_Video_Prompt), and an autonomous media routing tag (Output_4_Asset_Type).
  • Text Normalization: A Regular Expression (regexp:Replace) module cleans raw markdown artifacts (stripping ** tags) and standardizes newline characters to prevent broken formatting across social networks.
  • Dynamic Asset Generation & Routing:
  • Video Branch (gemini-omni-flash-preview): Generates 16:9 cinematic video. A secondary router distributes the asset to:
  1. Facebook Pages: Uploads the video file and caption, pauses for 120 seconds via a sleep utility to allow processing, and automatically appends a first comment containing direct resource links.
  2. YouTube: Publishes the video under the “Education” category with automated metadata tagging.
  3. LinkedIn: Creates a native video post directly on the main feed.
  4. Notion & Email Logging: Archives the published post title, YouTube URL, prompt, and date to a Notion CMS database and sends a completion email digest.
  • Image Branch (gemini-3.1-flash-image): Filters for image tags, renders a 2K 4:3 high-fidelity image, and multi-casts the asset across:
  1. Facebook Pages: Posts the image and caption, waits 8 seconds, and posts the CTA comment.
  2. LinkedIn: Publishes an image share post.
  3. Notion & Email Logging: Uploads the raw file via Gemini File API, logs the record to the Notion database, and delivers a completion alert.

Governance & Quality Controls (The Guardrails)

To maintain enterprise reliability and brand safety, the architecture enforces strict structural and execution guardrails:

  • Deterministic Schema Enforcement: Enforces a rigid JSON schema via Gemini’s responseSchema parameters, eliminating parsing errors, dropped fields, or conversational AI filler.
  • Negative Prompt Constraints: The system instructions explicitly mandate hard negative boundaries (such as banning specific off-brand visual themes like “Ringworlds” and prohibiting watermarks/text overlays).
  • Automated Data Sanitization: Regex cleaning strips unsupported markdown syntax before content reaches APIs, protecting typography and readability.
  • Fault Tolerance & Error Handling: Integrated builtin:Ignore error handlers on critical database modules (e.g., Notion logging nodes) prevent API timeout hiccups from interrupting live social distribution.
  • Audit Logging & Telemetry: Automated email webhooks fire at both the initiation and termination of the workflow, providing immediate visibility and execution tracking for every run.

Measurable Business Impact (The ROI)

Automating the end-to-end creative lifecycle delivered significant operational improvements:

Quantified Financial ROI:

The automation recovers over $16,000 in annual operational capacity, freeing technical and creative resources to focus entirely on high-tier product development and curriculum design.