n8n Workflow: AI-Powered Information Extraction and Jira Integration

This n8n workflow demonstrates a powerful combination of AI capabilities and traditional automation to extract structured information from text and integrate it with Jira. It leverages LangChain nodes for AI processing, including document loading, text splitting, embeddings, and an AI agent for information extraction, before potentially interacting with Jira.

What it does

This workflow is designed to process textual input, extract specific information using AI, and then, based on certain conditions, perform actions within Jira.

1. Triggers on a Schedule: The workflow starts on a predefined schedule, indicating it's designed for periodic execution rather than real-time event-driven processing.
2. Loads Data: It uses a "Default Data Loader" to ingest some form of document or text content.
3. Splits Text: The "Recursive Character Text Splitter" prepares the loaded text for AI processing by breaking it down into manageable chunks.
4. Generates Embeddings: "Embeddings OpenAI" creates vector embeddings of the text, which are numerical representations useful for semantic search and AI understanding.
5. Stores Vectors: The "Supabase Vector Store" is used to store and manage these text embeddings, likely for efficient retrieval or similarity searches.
6. Extracts Information with AI: An "AI Agent" and an "Information Extractor" (powered by an "OpenAI Chat Model") work together to intelligently parse the text and extract structured data based on defined criteria.
7. Filters and Processes Extracted Data:
   • "Remove Duplicates" ensures that only unique pieces of extracted information are processed further.
   • "Split Out" likely separates the extracted information into individual items.
   • "Sort" organizes the extracted items.
   • "Summarize" aggregates or condenses the processed information.
8. Conditional Logic: An "If" node evaluates a condition on the processed data.
9. Jira Integration (Conditional): If the condition in the "If" node evaluates to TRUE, the workflow proceeds to interact with Jira Software. The specific action (e.g., creating a ticket, updating an issue) is not detailed in the JSON but is implied by the presence of the Jira node.
10. No Operation (Conditional): If the condition in the "If" node evaluates to FALSE, the workflow takes no further action, indicated by the "No Operation, do nothing" node.
11. Data Manipulation: An "Edit Fields (Set)" node is present, suggesting that data is transformed or modified at some point in the workflow, likely before or after AI processing.

Prerequisites/Requirements

To use this workflow, you will need:

• n8n Instance: A running n8n instance.
• OpenAI API Key: Required for the "Embeddings OpenAI" and "OpenAI Chat Model" nodes, used by the AI Agent and Information Extractor.
• Supabase Account: Required for the "Supabase Vector Store" to store and retrieve text embeddings.
• Jira Software Account: Required if the workflow's conditional logic leads to interaction with Jira. You'll need appropriate credentials configured in n8n.
• Textual Data Source: The "Default Data Loader" implies a source of text that this workflow will process.

Setup/Usage

1. Import the Workflow: Download the provided JSON and import it into your n8n instance.
2. Configure Credentials:
   • Set up your OpenAI API Key credentials in n8n.
   • Configure your Supabase credentials (API URL and API Key) in n8n.
   • If you intend to use the Jira integration, set up your Jira Software credentials in n8n.
3. Customize Data Loader: Adjust the "Default Data Loader" node to point to your specific data source (e.g., a file, a URL, or previous node's output).
4. Define AI Agent and Information Extractor Logic:
   • Configure the "AI Agent" and "Information Extractor" nodes with the specific prompts, tools, and schemas required to extract the desired information from your text. This is crucial for the AI to understand what to look for.
   • Ensure the "OpenAI Chat Model" is configured with the correct model and settings.
5. Configure Conditional Logic: Modify the "If" node (ID 20) to define the conditions under which the Jira integration should be triggered.
6. Customize Jira Action: If the Jira path is taken, configure the "Jira Software" node (ID 77) with the specific action you want to perform (e.g., create an issue, update a field) and map the extracted data to the relevant Jira fields.
7. Set Schedule: Adjust the "Schedule Trigger" node (ID 839) to define how often you want the workflow to run.
8. Activate the Workflow: Save and activate the workflow.