Step 2: Connect Google Drive

Download the new file content from Google Drive so it can be indexed.

1. Add Download Drive File and set Operation to download.
2. Set File ID to {{ $json.id }} so the node downloads the file detected by Drive File Trigger.
3. Connect Drive File Trigger → Download Drive File.
4. Credential Required: Connect your Google Drive credentials.

Step 3: Set Up Pinecone Knowledge Indexing

Split and load the file data, generate embeddings, and store vectors in Pinecone for retrieval.

1. Add Recursive Text Chunker and set Chunk Size to 500 and Chunk Overlap to 20.
2. Add Binary Data Loader and set Data Type to binary.
3. Add Pinecone Index Writer and set Mode to insert.
4. In Pinecone Index Writer, set Pinecone Namespace to Q&A and Pinecone Index to pineconeIndex.
5. Connect Recursive Text Chunker → Binary Data Loader → Pinecone Index Writer, and connect OpenAI Embedding Maker as the embedding model for Pinecone Index Writer.
6. Credential Required: Connect your Pinecone credentials.

Step 4: Set Up Conversational AI Agent

Configure the AI agent, memory, and company knowledge retrieval so the chat assistant can answer questions and collect lead details.

1. Open Conversational AI Agent and set System Message to =## Role: You are a friendly assistant for a Company named *New Company**. ## Task: You answer questions about the business. ## Details: You have access to various tools, which you use correctly. ## Tools: - **Company Knowledge Tool** Use this tool to answer questions with knowledge about the company. - **Append Lead Row** Use this tool to store contact information such as name, email, interestet in and phone number. After a customer asks about opening hours, products, location, or business information, ask them for their name, email, specific interests and phone number. .
2. Connect Primary Chat Model to Conversational AI Agent and set the model to gpt-4o.
3. Connect Conversation Window Memory and set Context Window Length to 12.
4. Configure Company Knowledge Tool with description gives answers related to the company newCompany, and connect Pinecone Index Reader with Pinecone Namespace set to Q&A and Pinecone Index to =pineconeIndex.
5. Connect Embedding Generator 2 as the embedding model for Pinecone Index Reader, and connect Secondary Chat Model (model gpt-4o) as the language model for Company Knowledge Tool.
6. Credential Required: Connect your OpenAI and Pinecone credentials.

Step 5: Configure Lead Capture Output

Store qualified lead details in Google Sheets using the AI tool output.

1. Open Append Lead Row and set Operation to append.
2. Map the columns to AI outputs:
   Name → {{ /*n8n-auto-generated-fromAI-override*/ $fromAI('Name', ``, 'string') }}
   Email → {{ /*n8n-auto-generated-fromAI-override*/ $fromAI('Email', ``, 'string') }}
   Phone → {{ /*n8n-auto-generated-fromAI-override*/ $fromAI('Phone', ``, 'string') }}
   Interestet in → {{ /*n8n-auto-generated-fromAI-override*/ $fromAI('Interestet_in', ``, 'string') }}.
3. Select your Google Sheet by setting Document ID and Sheet Name (currently empty in the workflow).
4. Credential Required: Connect your Google Sheets credentials.

Step 6: Test and Activate Your Workflow

Validate both ingestion and chat flows before enabling the workflow.

1. Manually execute the workflow and upload a file into the watched Google Drive folder to confirm Drive File Trigger → Download Drive File → Pinecone Index Writer completes without errors.
2. Send a test message to the Chat Message Webhook URL and confirm Conversational AI Agent responds with company information.
3. Provide name, email, phone, and interest in the chat and verify a new row is appended by Append Lead Row.
4. Once verified, toggle the workflow to Active to run it in production.