Overview

I built a LinkedIn content automation engine that converts TLDR AI newsletters into SOC-focused LinkedIn posts with attached Gamma slide decks—on a recurring schedule—using Make.com, OpenRouter, Gamma AI, and HTTP modules. The automation ingests newsletters from Gmail, extracts all articles, decides which one is most relevant for SOC/detection engineering, pulls the full article content, generates a LinkedIn post in my voice, creates a Gamma slideshow, converts the export link into a downloadable file URL, then publishes a LinkedIn post containing the text and infographic link. Why this matters:
Security professionals need consistent, high-signal content without spending hours drafting posts. This workflow shows how to combine AI + automation to maintain a strong technical presence on LinkedIn while staying focused on real SOC work.

Key Features

• Gmail-driven newsletter ingestion:
  • A Gmail Watch Emails trigger listens for incoming TLDR newsletters based on sender/subject filters.
  • Each new newsletter kicks off a single pass through the pipeline so posts are always based on the latest content.
• Article extraction and decision logic:
  • An OpenRouter “Extractor” step parses the newsletter body and returns all articles, links, and type metadata (e.g., AI, security, etc.).
  • A Parse JSON module structures this output into arrays/objects Make can work with.
  • A second OpenRouter “Article Decider” call ranks the extracted articles and selects the one best aligned with SOC, detection engineering, or security automation.
  • Another Parse JSON module converts the decider’s response into clean fields: chosen title, URL, and summary.
• Full article retrieval and post generation:
  • An HTTP Get a web page module downloads the full content of the selected article using its URL.
  • An OpenRouter “LinkedIn Post Text” step takes the article (title, link, content) plus a style guide prompt and generates a LinkedIn-ready post in my voice as a SOC analyst.
  • The prompt enforces:
    • 🧠 hook,
    • 🔹 technical bullets,
    • ⚠️ defender takeaway,
    • source link + security hashtags.
• Gamma slideshow creation and export handling:
  • A Gamma module generates a multi-slide slideshow/infographic using the chosen article and a structured prompt (slide structure, headings, SOC insights).
  • Gamma returns a PDF export URL for the finished deck.
  • An HTTP module converts that export URL into a direct downloadable link that can be safely shared in the LinkedIn post text.
• LinkedIn posting & scheduling:
  • A final LinkedIn Create User Text Post module publishes the AI-generated text and appends the Gamma deck link so viewers can open and scroll all slides.
  • A lightweight Sleep step can be used as a buffer (e.g., giving Gamma time to finish) but the flow works even without it.
  • The entire scenario is scheduled in Make.com to run every Tuesday, Thursday, and Saturday at 9:58 AM ET, aligning with a consistent posting cadence.

Technologies Used

• Automation Platform: Make.com (scenario orchestration, scheduling, JSON parsing, variable mapping).
• LLM / AI APIs: OpenRouter (article extraction, ranking, LinkedIn copy generation).
• Presentation / Infographic Engine: Gamma AI (slide deck generation, PDF export links).
• Email Integration: Gmail Watch Emails (newsletter ingestion).
• Web Content Retrieval: HTTP modules (fetch article body, handle Gamma export URLs).
• Social Platform: LinkedIn (Create User Text Post with embedded infographic link).
• Data Structures: JSON parsing, arrays/objects, prompt templates, and mapped variables across modules.

Challenges and Learnings

• Chaining AI steps reliably:
  • Needed to carefully design prompts so the Extractor, Article Decider, and Post Text models each returned predictable JSON and text formats.
  • Learned that enforcing explicit JSON schemas and word limits keeps the automation from breaking when newsletter formatting changes.
• Handling external content and timing:
  • Some articles load slowly or have heavy pages, so the HTTP fetch and Gamma generation steps were tested repeatedly with different content to make sure they wouldn’t time out.
  • Added an optional Sleep module as a safety buffer after Gamma when experimenting, then tuned the flow once everything proved stable.
• Balancing depth and readability:
  • Iterated on the LinkedIn prompt so posts stay technical enough for SOC analysts but short and scannable for a feed environment.
  • Adjusted bullets, hooks, and hashtag count to land in a sweet spot for engagement and clarity.

Takeaway:
The project shows how to turn a raw information stream (newsletters) into production-ready, security-focused content using layered LLM calls and automation. The key is structuring each step so the next module always knows exactly what to expect.

Outcome

• Consistent, automated posting:
  • The engine generates and posts new SOC/detection-focused content three times per week at 9:58 AM ET without manual intervention.
• High-leverage personal branding:
  • Each run produces:
    • 1 curated article chosen for security relevance.
    • 1 LinkedIn post tailored to SOC analysts and detection engineers.
    • 1 Gamma slideshow link that functions as a visual infographic.
• Reusable framework:
  • The same pattern (email → extraction → decision → content → visual → post) can be reused for other sources (blogs, RSS, internal reports) and other platforms (e.g., X/Twitter, newsletters).
  • Demonstrates practical skills in Make.com scenario design, multi-LLM orchestration, API integration, and security-focused content automation.