Qualify B2B Prospects with ProspectPro Web RAG and GPT-4

This n8n workflow streamlines the process of qualifying B2B prospects by leveraging web data retrieval (RAG - Retrieval Augmented Generation) and advanced AI capabilities (GPT-4). It takes a prospect's company name as input, enriches it with web-scraped information, and then uses a large language model to determine if the prospect is a good fit based on predefined criteria.

What it does

This workflow automates the following steps:

1. Receives a Company Name: It is designed to be triggered by another workflow, expecting a company name as input.
2. Enriches Company Data (Planned): The workflow is structured to incorporate a web scraping/RAG step (currently represented by a placeholder) to gather relevant information about the company from the web.
3. Analyzes Prospect Fit with AI: It uses an OpenAI Chat Model (GPT-4) with a structured output parser to evaluate the prospect based on the gathered information and specific qualification criteria.
4. Routes Based on Qualification: An 'If' node then checks the AI's qualification output to determine if the prospect is a "good fit" or not.
5. Performs Conditional Actions (Planned): Depending on whether the prospect is qualified, the workflow is designed to branch out to further actions (e.g., add to CRM, send notification, etc. - these are not explicitly defined in the provided JSON but are implied by the 'If' node).

Prerequisites/Requirements

• n8n Instance: A running n8n instance.
• OpenAI API Key: An API key for OpenAI to use the GPT-4 chat model. This will need to be configured as an n8n credential.
• Web Scraping/RAG Tool (Planned): While not explicitly defined in the JSON, the workflow's purpose implies the need for a web scraping or RAG (Retrieval Augmented Generation) tool/service to gather company data. This would likely be implemented using additional n8n nodes (e.g., HTTP Request, Browser, or a specialized RAG integration).

Setup/Usage

1. Import the Workflow:
   • Download the provided JSON.
   • In your n8n instance, go to "Workflows" and click "New".
   • Click the three dots menu (⋮) next to the workflow name and select "Import from JSON".
   • Paste the workflow JSON or upload the file.
2. Configure Credentials:
   • Locate the "OpenAI Chat Model" node.
   • Click on the "Credential" field and select an existing OpenAI credential or create a new one, providing your OpenAI API Key.
3. Implement Web Data Retrieval:
   • The "Code" node currently acts as a placeholder for data enrichment. You will need to replace or augment this with nodes that perform web scraping or integrate with a RAG service (e.g., using an HTTP Request node to an external API, or a Browser node for direct scraping).
   • Ensure the output of your data retrieval step provides the necessary context for the "Basic LLM Chain" and "OpenAI Chat Model" to evaluate the prospect.
4. Define Qualification Logic:
   • Review the "Basic LLM Chain" and "OpenAI Chat Model" nodes. The prompt within these nodes should clearly define what constitutes a "good fit" for a B2B prospect.
   • The "Structured Output Parser" expects a specific JSON schema for the AI's response (e.g., { "is_qualified": boolean, "reason": string }). Ensure your LLM prompt guides the AI to output in this format.
   • Configure the "If" node (ID 20) to evaluate the output from the "Structured Output Parser" (e.g., {{ $json.is_qualified }} equals true).
5. Add Post-Qualification Actions:
   • Connect nodes to the "True" and "False" branches of the "If" node (ID 20) to handle qualified and unqualified prospects accordingly. Examples include:
     • Adding qualified prospects to a CRM (e.g., HubSpot, Salesforce, Pipedrive).
     • Sending notifications (e.g., Slack, Email) for qualified leads.
     • Logging unqualified prospects for review.
6. Activate the Workflow: Once configured, activate the workflow. It will then be ready to be triggered by another workflow.

Note: The current JSON does not include the actual web scraping or RAG implementation, nor does it define the specific actions for qualified/unqualified prospects. These steps are crucial for a fully functional workflow and need to be added based on your specific requirements.