Step 2: Connect Google Sheets

Configure the three append operations that log each inquiry type to Google Sheets.

1. Open Append Sheet - Product and set Operation to append.
2. Set Document to [YOUR_ID] and Sheet to gid=0 (name: Foglio1).
3. Map DATE to {{ $now.format('dd/MM/yyyy') }} and DRAFT to {{ $('Compose Reply - Product').item.json.output.text }}.
4. Map FIRST NAME, LAST NAME, EMAIL , PHONE, and MESSAGE to the corresponding Map Contact Fields expressions.
5. Repeat the same configuration for Append Sheet - Order and Append Sheet - Other, using their respective {{ $('Compose Reply - Order').item.json.output.text }} and {{ $('Compose Reply - Other').item.json.output.text }} expressions.
6. Credential Required: Connect your googleSheetsOAuth2Api credentials to Append Sheet - Product, Append Sheet - Order, and Append Sheet - Other.

Step 3: Set Up AI Classification and Field Mapping

Normalize inbound data and classify the inquiry so it routes to the correct AI response path.

1. Open Map Contact Fields and set assignments:
   • first_name → {{ $json.body.first_name }}
   • last_name → {{ $json.body.last_name }}
   • email → {{ $json.body.email }}
   • phone → {{ $json.body.phone }}
   • message → {{ $json.body.message }}
2. Open Route Inquiry Types and set Input Text to {{ $json.message }}.
3. Verify categories in Route Inquiry Types include Product Info and Order Info, with Fallback set to other.
4. Ensure Gemini Classifier Model is connected as the language model for Route Inquiry Types with Model Name models/gemini-2.0-flash.
5. Credential Required: Connect your googlePalmApi credentials to Gemini Classifier Model.

Step 4: Set Up AI Reply Generation

Generate tailored draft responses for product, order, and other inquiries.

1. Open Compose Reply - Product and set Text to the multi-line prompt using contact fields, starting with This is the message you received that you need to reply to: and including {{ $('Map Contact Fields').item.json.first_name }}, {{ $('Map Contact Fields').item.json.last_name }}, {{ $('Map Contact Fields').item.json.email }}, {{ $('Map Contact Fields').item.json.phone }}, and {{ $('Map Contact Fields').item.json.message }}.
2. Repeat the same Text and promptType define configuration for Compose Reply - Order and Compose Reply - Other.
3. Confirm Gemini Reply Model C is the language model for Compose Reply - Product, Gemini Reply Model B for Compose Reply - Order, and Gemini Reply Model A for Compose Reply - Other, each using models/gemini-2.0-flash-exp.
4. Credential Required: Connect your googlePalmApi credentials to Gemini Reply Model A, Gemini Reply Model B, and Gemini Reply Model C.
5. Keep Parse Subject/Text A, Parse Subject/Text B, and Parse Subject/Text C attached as output parsers with their Input Schema for subject and text. These are AI sub-nodes—credentials should be added to the parent Gemini nodes, not the parsers.

Step 5: Configure Draft Email Outputs

Create Gmail drafts using the generated subject and message from the AI outputs.

1. Open Draft Email - Product and set Resource to draft.
2. Set Subject to {{ $json.output.subject }} and Message to the multi-line template using {{ $json.output.text }} plus the mapped contact fields.
3. Set Send To to {{ $('Route Inquiry Types').item.json.email }} so the draft targets the detected email address.
4. Repeat the same configuration for Draft Email - Order and Draft Email - Other.
5. Credential Required: Connect your gmailOAuth2 credentials to Draft Email - Product, Draft Email - Order, and Draft Email - Other.

Step 6: Test and Activate Your Workflow

Run a full test from the webhook to confirm drafts and sheet logs are created correctly.

1. Click Execute Workflow and submit a test form to the webhook URL from Incoming WP Form Hook.
2. Confirm Map Contact Fields outputs the expected fields and Route Inquiry Types classifies the message.
3. Verify a Gmail draft is created in the appropriate node: Draft Email - Product, Draft Email - Order, or Draft Email - Other.
4. Check that a new row was appended in the corresponding sheet node (Append Sheet - Product, Append Sheet - Order, or Append Sheet - Other) with the AI draft and contact data.
5. When tests pass, toggle the workflow to Active for production use.