Contact form messages routed by AI and logged to Sheets

Contact form messages routed by AI and logged to Sheets

An n8n workflow with a contact form for an online shop. A Text Classifier sorts each message as a quote request, product information, a general problem, an order question or other. The message is then emailed to the matching department and added as a row to a Google Sheet.

What is inside

• workflow.json: the workflow as n8n exports it, fourteen nodes: thirteen that do the work and one sticky note. It holds no credential.
• README.md: these notes.

What we changed from the original copy

• The credential links are gone. Select your own for OpenAI, SMTP and Google Sheets.
• The model was gpt-4o-mini. OpenAI's deprecations page, read on 12 November 2026, gives 14 January 2027 as its shutdown date and names gpt-6-luna.
• The Text Classifier reads the model's reply as JSON text and does not call a tool, as n8n's source shows. The OpenAI node therefore names gpt-6-luna.
• The Product info output led nowhere, and the Request Quote output fed the product department as well as the quote department. Request Quote now goes to Quote Dep. only, and Product info goes to Prod. Dep..
• All five Google Sheets nodes wrote the same value, info prodotti, in CATEGORIA. Each one now writes its own category: Request Quote, Product info, General problem, Order or other.
• The email body ended with a Tipo prodotto line that reads a field the form does not have. That line is removed.
• The addresses and the sheet id are placeholders: <YOUR_QUOTES_EMAIL>, <YOUR_PRODUCTS_EMAIL>, <YOUR_SUPPORT_EMAIL>, <YOUR_ORDERS_EMAIL>, <YOUR_OTHER_EMAIL>, <YOUR_SENDER_EMAIL> and <YOUR_SHEET_ID>. The tab is named Sheet1.
• The note is rewritten, with a setup list.
• Nothing else.

Before you start

• An n8n instance, on n8n Cloud or your own server.
• An OpenAI API key.
• An SMTP account to send from, and the address it sends as.
• A Google Sheet whose first row holds the headers DATA, NOME, EMAIL, RICHIESTA, CATEGORIA and TO.

Set it up

1. In n8n, create a workflow, open its menu and choose Import from File. Pick workflow.json.
2. Add an OpenAI credential and select it in OpenAI.
3. Add an SMTP credential and select it in the five department nodes. Replace each <YOUR_..._EMAIL> placeholder with the address that should receive that kind of message.
4. Add a Google Sheets credential and select it in the five database nodes. Replace <YOUR_SHEET_ID> in each one with your spreadsheet's id and check the tab name.
5. Open the form's test address from On form submission, send a message and look at the inbox and the sheet.

How a run goes

1. On form submission shows a form titled Contacts with the required fields Name, Email and Message. It answers the visitor when the last node finishes.
2. Text Classifier reads the message and picks one of four categories, or the fallback other.
3. One of the five department nodes sends an email. The subject starts with [n8n Contacts] and the department, and the reply address is the visitor's.
4. The matching database node adds a row. DATA is the submission time, NOME, EMAIL and RICHIESTA come from the form, CATEGORIA is the category and TO is the recipient.

Change the look

• The categories and their descriptions: Text Classifier. A new category adds an output, which needs its own email and sheet nodes.
• Who gets which message: toEmail in each department node.
• The email wording: the body in each department node.
• The form's fields: On form submission.

Good to know

• The workflow sends real email, so use your own addresses while you test.
• A message that fits no category goes to Other Dep., because the fallback option is set to other.
• This copy was not run against OpenAI, an SMTP server or Google Sheets.
• The message text goes to OpenAI. The name, address and message go to the departments and the sheet.
• The JSON holds no credential.