n8n Workflow: Automate Competitor Research with AI Agents and Notion

This n8n workflow leverages AI agents to perform competitor research by fetching data from a specified API, processing it, and then organizing the findings into a Notion database. It streamlines the process of gathering and structuring competitive intelligence.

What it does

This workflow automates the following steps:

1. Manual Trigger: Initiates the workflow manually, allowing you to start the competitor research on demand.
2. HTTP Request: Makes an API call to an external service (likely Exa AI or a similar data source) to fetch raw competitor data.
3. Edit Fields (Set): Transforms and prepares the data received from the HTTP request, potentially extracting relevant fields or renaming them for consistency.
4. Loop Over Items (Split in Batches): Processes the fetched data in batches, iterating through each competitor entry or data point.
5. Remove Duplicates: Ensures data quality by removing any duplicate entries within the processed batch, preventing redundant information.
6. Limit: Controls the number of items processed in each batch, useful for managing API rate limits or processing large datasets incrementally.
7. Split Out: Further refines the data by splitting out nested arrays or objects into individual items, making them easier for the AI agent to process.
8. AI Agent (Langchain): Utilizes an AI agent powered by Langchain to analyze the competitor data. This agent likely performs tasks such as:
   • Summarizing competitor profiles.
   • Identifying key strengths and weaknesses.
   • Extracting specific metrics or features.
   • Structuring the information according to a predefined schema.
9. OpenAI Chat Model: Interacts with the OpenAI Chat API as the underlying language model for the AI Agent, enabling advanced natural language understanding and generation capabilities.
10. Structured Output Parser: Parses the AI agent's output into a structured format (e.g., JSON), ensuring that the insights are consistently formatted and ready for database insertion.
11. Wait: Introduces a delay in the workflow, which can be useful for respecting API rate limits or allowing time for external services to process requests.
12. Notion: Creates new items or updates existing ones in a specified Notion database with the structured competitor research findings.

Prerequisites/Requirements

To use this workflow, you will need:

• n8n Instance: A running n8n instance (cloud or self-hosted).
• API Endpoint for Competitor Data: Access to an external API (e.g., Exa AI) that provides competitor information. You will need the API URL and any necessary authentication.
• OpenAI API Key: An OpenAI API key for the "OpenAI Chat Model" node.
• Notion Account: A Notion account with a pre-configured database where the competitor research data will be stored. You will need the database ID and a Notion integration token.

Setup/Usage

1. Import the Workflow:
   • Download the provided JSON file.
   • In your n8n instance, click on "Workflows" in the left sidebar.
   • Click "New" -> "Import from JSON" and paste the workflow JSON or upload the file.
2. Configure Credentials:
   • HTTP Request: Update the "HTTP Request" node with the URL of your competitor data API and any required authentication (e.g., API keys in headers or query parameters).
   • OpenAI Chat Model: Configure the "OpenAI Chat Model" node with your OpenAI API key.
   • Notion: Set up a Notion credential in n8n and configure the "Notion" node with your Notion database ID and the appropriate properties mapping.
3. Customize AI Agent:
   • Review the "AI Agent" node's configuration. You may need to adjust the prompt and tools used by the agent to tailor the research to your specific needs.
   • The "Structured Output Parser" node defines the expected output structure from the AI. Ensure this matches your Notion database schema.
4. Activate and Execute:
   • Once configured, activate the workflow.
   • You can execute the workflow manually by clicking "Execute Workflow" on the "Manual Trigger" node.
   • For continuous monitoring, you might consider replacing the "Manual Trigger" with a schedule-based trigger (e.g., "Cron" node) to run the research periodically.