Automated Google Meet Analysis and Task Creation - n8n Workflow

Project Managers needing immediate, structured summaries of recorded meetings.
Technical Teams looking for advanced examples of using AssemblyAI and Anthropic Claude within an n8n workflow.
Operations Staff who want to automate task creation from verbal agreements.
Anyone seeking robust, AI-driven n8n templates for productivity.

Managing meeting follow-ups and ensuring action items are captured can be time-consuming. This intelligent n8n workflow solves this problem by completely automating the post-meeting process for Google Meet recordings. It integrates seamlessly with Google Calendar and Google Drive to locate the meeting recording. Using the AssemblyAI n8n node, it generates a high-quality transcript with speaker separation. Finally, the Anthropic Claude AI (via the Langchain n8n node) analyzes the transcript, extracting concise summaries, detailed action items, and overall sentiment. The resulting insights are then distributed instantly to Slack and Notion, transforming raw recording data into structured, actionable tasks. This comprehensive n8n solution ensures no commitment is missed and drastically improves team accountability.

This powerful n8n workflow executes through a series of chained nodes:

1. Trigger & Fetch: The process starts manually (or can be adapted to an n8n trigger like a schedule or webhook). The workflow queries Google Calendar for confirmed Google Meet events that occurred in the last 24 hours.


2. Recording Retrieval: It filters for valid Google Meet entries and searches Google Drive for the corresponding recording file. If no recording is found, the n8n workflow stops with an error.


3. Transcription: The recording is downloaded via the Google Drive n8n node and then uploaded to AssemblyAI. A subsequent HTTP Request n8n node initiates and polls the transcription process until completion, ensuring speaker labels and highlights are included.


4. AI Analysis: The formatted transcript is passed to the specialized Langchain Agent n8n node, which utilizes the Claude 3.5 Sonnet model. The prompt instructs the AI to extract a summary, action items (with assignees and dates), key dates, and meeting sentiment.


5. Data Structuring: A Code n8n node meticulously parses the structured text output from the AI into a clean JSON object.


6. Distribution: The full summary is immediately posted to Slack. If the AI identified action items, the n8n workflow uses the Split Out n8n node to process each item individually.


7. Task Creation: For every distinct action item, a Notion n8n node creates a dedicated task entry, complete with the task title, assignee, and due date, ensuring immediate task management.

To implement this n8n templates workflow, follow these steps:

1. Import: Copy the provided JSON and import it into your n8n instance using the 'New' -> 'Import from JSON' option.


2. Credentials: You will need to set up and connect the following credentials:

1. Customization: Review the postToSlack n8n node and update the CHANNEL_ID parameter. Also, ensure the Notion n8n node targets the correct database URL/ID.


2. Testing: Click the whenClickingTestWorkflow n8n trigger to run the flow against your recent meetings.

whenClickingTestWorkflow (Manual Trigger): The starting n8n trigger for testing. In production, this can be replaced by a Schedule n8n node.
getRecentMeetings (Google Calendar n8n node): Fetches calendar events. Key Configuration: Looks back 24 hours (timeMin) and lists up to 10 events.
filterGoogleMeetEvents (Filter n8n node): Filters results to only confirmed events where the conferenceSolution.name is exactly 'Google Meet'.
downloadRecording (Google Drive n8n node): Downloads the file found in the previous search step. Key Configuration: Uses the file ID from the preceding Google Drive n8n node output.
uploadToAssemblyAI (HTTP Request n8n node): Uploads the binary file data to the AssemblyAI API endpoint for preparation.
requestTranscription (HTTP Request n8n node): Initiates transcription with advanced features. Key Configuration: POST request to /v2/transcript, enabling speakerlabels and autohighlights.
pollTranscriptionStatus (HTTP Request n8n node): Continuously checks the transcription status. Key Configuration: Uses the ID of the pending job in the URL. Timeout set to 5 minutes (300000ms).
analyzeMeetingWithAI (Langchain Agent n8n node): The core AI processing step. Key Configuration: Uses Anthropic's Claude 3.5 Sonnet model (claudeChatModel) to analyze the transcript based on a detailed instruction prompt to extract summary and action items.
parseAIOutput (Code n8n node): Crucial for transforming the free-text AI response into structured actionItems, summary, and sentiment JSON fields for subsequent n8n nodes.
postToSlack (Slack n8n node): Publishes the final analysis to the team channel.
splitActionItems (Split Out n8n node): Prepares the array of action items for individual task creation.
createNotionTask (Notion n8n node): Creates individual tasks based on the data provided by the preceding Split Out n8n node.