AI Sales Call Summarization and Queue Processor for Notion - n8n Workflow

Sales Operations and RevOps teams needing structured data from Gong calls.
Technical users looking for advanced, resilient n8n templates for queue processing.
Developers implementing multi-step AI agents using a core n8n node orchestration.
Anyone needing to ensure only new items are processed in a large automation batch.

This comprehensive n8n workflow, often referred to as CallForge, is designed to bring resilience and structure to high-volume data processing tasks, specifically handling sales call records. The core problem it solves is efficiently managing a queue of new calls, preventing duplicates, and ensuring that complex AI analysis—which often involves API rate limits—can be resumed even if the process fails mid-run. This n8n workflow acts as the master queue processor. It starts by checking the existing call records in Notion to identify only new calls. It then creates the parent object in Notion for each new call, capturing crucial metadata like Salesforce Opportunity IDs and attendee lists. Finally, it passes the data and the new Notion page ID to an external AI sub-workflow for deep analysis, while simultaneously providing real-time progress updates via the Slack n8n node. Utilizing this robust n8n template dramatically improves data reliability and operational visibility.

The operational flow of this n8n workflow focuses on resilience, deduplication, and sequential processing:

1. Trigger and Preparation: The process starts with the Execute Workflow Trigger n8n node, receiving the raw sales call data (likely from a preceding Gong collector n8n workflow). This data is aggregated by the Reduce down to 1 object n8n node.


2. Deduplication: The n8n workflow uses the Get all older Calls Notion n8n node to retrieve all existing call IDs stored in the Notion database. The Only Process New Calls n8n node (a Compare Datasets type) performs robust deduplication, ensuring only calls not already in Notion proceed.


3. Job Initiation Alert: The Post Slack Receipt n8n node immediately alerts the team that the processing job has begun, noting the exact number of new calls found. The resulting Slack message data is saved to allow for progress updates later.


4. Loop Setup: The new, unique call data, merged with the Slack alert details, enters the Loop Over Calls n8n node (Split In Batches), which processes each call linearly.


5. Notion Parent Creation: For each item, the Create Notion DB Page n8n node establishes the parent call record in the database, populating fields with metadata like SF data and attendee names. This step generates a crucial Notion page ID.


6. AI Orchestration: The call data and the new Notion ID are merged and passed to the AI Team Processor n8n node (an Execute Workflow node). This sub-workflow handles the complex AI analysis, using the Notion Parent ID to structure the results.


7. Progress Tracking: The Update Slack Progress n8n node uses the stored Slack message receipt to dynamically update the original alert, providing real-time visibility into the queue's progress before looping back to the next call via the Loop to next call n8n node.


8. Job Completion: Once the loop finishes processing all calls, the final items are bundled, and the Post Completed Calls Message Slack n8n node sends a final notification confirming the successful conclusion of the n8n workflow execution.

To deploy this powerful n8n workflow, follow these setup instructions:

1. Import: Copy the provided JSON data and paste it directly into your n8n instance via the 'Import Workflow' option.


2. Credentials: You must configure credentials for the necessary n8n node services:

1. Sub-Workflow Dependency: Ensure the linked sub-workflow, named 'AI Team Processor Demo' (ID 4Uol9xlNKyNH213f), is also imported and active in your n8n instance, as this n8n workflow depends on it for the heavy lifting of AI analysis.


2. Test: Execute the Execute Workflow Trigger n8n trigger manually with sample data mirroring your Gong call records to verify all connections and data transformations are functioning correctly before production use.

Execute Workflow Trigger (n8n trigger): This initial n8n node serves as the entry point, receiving the structured list of sales call data items that need processing.
Notion (Get all older Calls): This n8n node is vital for deduplication. It connects to the Notion database to fetch all existing records, specifically looking for previously processed call IDs.
Compare Datasets (Only Process New Calls): A core n8n node logic tool. It compares the incoming call data list with the list of existing Call IDs from Notion, allowing the n8n workflow to proceed only with items that are genuinely new.
Slack (Post Slack Receipt / Update Slack Progress / Post Completed Calls Message): These dedicated Slack n8n nodes handle communication. The first posts a starting message; the second updates that message in real-time within the processing loop using the message's ts (timestamp) property; and the third confirms job completion. This utilizes a powerful feature of the Slack n8n node.
Split In Batches (Loop Over Calls): This n8n node ensures that the massive list of calls is processed one by one (a batch size of 1), which is crucial for managing rate limits associated with the Notion n8n node and the subsequent AI services.
Notion (Create Notion DB Page): This n8n node creates the parent database entry, mapping rich contextual data (Salesforce links, dates, and attendees) into the Notion database properties. The resulting Notion page ID is key for linking subsequent AI-generated data.

• Execute Workflow (AI Team Processor): This is the bridge to the AI logic. This n8n node calls a separate, modular n8n workflow template designed to run specialized AI analysis tasks on the call transcript, passing both the call data and the Notion parent ID for context.