AI Gong Sales Call Data Preprocessor and Deduplicator - n8n Workflow

• Sales Operations teams needing automated analysis of Gong call transcripts.


• Data scientists or developers looking for advanced n8n templates for complex data preparation.


• Users who need to combine data from multiple sources (Gong, Notion, Google Sheets) before running an AI model.


• Anyone utilizing n8n to build modular, efficient data pipelines that prevent duplicate processing.

This comprehensive automation is an essential component for high-volume sales analysis systems. It solves the critical problem of preparing raw, heterogeneous data for effective AI consumption. Starting from an incoming n8n trigger, this n8n workflow orchestrates data retrieval from three different platforms: Gong (for calls), Google Sheets (for product integrations), and Notion (for competitor lists).

The core value lies in data enrichment and robust deduplication. By aggregating supporting information (competitors, integrations) and injecting it into the prompt context, the AI processing step is significantly enhanced. The workflow uses sophisticated n8n node logic—specifically the Compare Datasets n8n node—to check new Gong calls against a log of previously processed calls stored in Notion. This ensures that only novel data is passed downstream, saving processing time and reducing costs associated with repeat AI summarization tasks. This powerful n8n workflow structure is typical of modular n8n templates designed for scale.

The entire process is initiated by an Execute Workflow Trigger n8n trigger, often called by a parent scheduler n8n workflow or webhook.

1. Parallel Data Fetching: The n8n workflow immediately branches out to fetch recent Gong calls, retrieve product integrations from Google Sheets (Get Integrations), and pull competitor data from a Notion database (Get list of Competitors).


2. Data Transformation & Enrichment: Aggregate and Set n8n node operations convert the integration and competitor lists into clean, comma-separated strings for optimal inclusion in the AI prompt.


3. Central Aggregation: All three data streams (Gong calls, integrations, competitors) are brought together using Merge 3 objects into one and subsequent Aggregate n8n node steps.


4. Deduplication Setup: The n8n workflow simultaneously queries a Notion database (Get Previous Phone Calls) containing the IDs of all calls already summarized. The IDs are isolated using an Isolate Only Call IDs n8n node.


5. Duplicate Removal: The enriched call data and the historical IDs are fed into the crucial Only Process New Calls n8n node (Compare Datasets). This n8n node eliminates any call that has been previously processed, allowing only unique, new items to proceed.


6. Batch Processing: The remaining unique calls are processed sequentially using the Loop Over Calls n8n node (Split In Batches).


7. Modular AI Processing: Each transcript is first passed to a nested n8n workflow (Transcript Processor) via an Execute Workflow n8n node for clean-up. The cleaned transcript is then passed to the main AI processing n8n workflow (Call Processor Demo) for structured summarization, completing the action of this powerful n8n workflow.

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


2. Set Credentials: This n8n workflow requires credentials for:

1. Configure Sub-workflows: Ensure the two linked external n8n templates/workflows (Transcript Processor and Call Processor Demo) are installed and accessible by their respective IDs (7BAQDjnHQVYO1SWG and cg4Eo7yZlhWkqHCB). Update the Execute Workflow n8n node IDs if necessary.


2. Database Configuration: Verify the database IDs in the Get list of Competitors and Get Previous Phone Calls n8n node match your specific Notion databases for competitors and processed call logs.

Execute Workflow Trigger: The starting n8n trigger for the flow, designed to be called by a master workflow or scheduler.
Gong: Fetches call data from the last two days. Crucial n8n node for source data capture.
Get Integrations (Google Sheets): Reads a list of integrations from a specific Google Sheets document, providing enrichment data for the AI prompt.
Get list of Competitors (Notion): Retrieves competitor names from a Notion database property, adding context for the AI model.
Comma Separate Integrations/Competitors (Set n8n node): Transforms array data from Google Sheets and Notion into single, comma-separated strings suitable for embedding into an AI system prompt.
Merge 3 objects into one / Aggregate n8n node: Core logic nodes responsible for combining the Gong call items with the global enrichment data (Integrations and Competitors).
Get Previous Phone Calls (Notion): Queries the historical Notion database where finalized summaries are stored, enabling the deduplication process.
Only Process New Calls (Compare Datasets n8n node): This key comparison n8n node analyzes incoming Gong IDs against historical Notion IDs, ensuring only new calls remain in the n8n workflow queue.
Loop Over Calls (Split In Batches n8n node): Iterates through the deduplicated list of calls, ensuring each transcript is processed individually.
Transcript Processor (Execute Workflow n8n node): Calls an external n8n workflow (a common modular design in advanced n8n templates) to clean and prepare the raw transcript text.

• Process All Call Transcripts (Execute Workflow n8n node): The final step, passing the enriched and cleaned data to the core AI summarizing n8n workflow.