You move from planning into building. Your goal is to assemble the first working version of the multimedia content creator and verify that the logic and AI outputs behave as expected.

Initial Setup

Before opening Make.com:

• Create a Google Sheet with columns: Source link, Source title, Tweet, Blog post, LinkedIn post;
• Create a Google Drive folder for future documents.

Build the RSS Input

Create a new scenario and add Retrieve RSS Feed Items as the first module.

Generate an XML feed using a tool like RSS.app. Use a broad topic such as tech to ensure AI-related headlines appear. Paste the feed URL into Make, set the maximum items to 10, and save the module.

Filter for AI Headlines

Add a router after the RSS module and run the scenario once to load sample data. Create a filter using:

• Field: Title;
• Operator: contains (case sensitive);
• Value: AI.

Create the AI Agents

You need three AI agents:

• Tweet writer;
• LinkedIn post writer;
• Blog writer targeting around 500 words.

Prepare prompts using an LLM. Each agent uses the same inputs (title, description, summary, URL) but produces a different format. Create each agent in Make using the largest available model. No tools or MCP are required.

Connect Router to Agents

From the router, create three branches and apply the same AI filter to each one. Add a Run an Agent module on every branch and map:

• Title;
• Description;
• Summary;
• URL.

Test the Scenario

Run the scenario and confirm. Once verified, the logic layer is complete.

• Only headlines with uppercase AI pass through;
• Each passing item triggers three AI outputs;
• Output length and structure match the platform.