Cold emails and follow-ups written by GPT for a lead list

Cold emails and follow-ups written by GPT for a lead list

An n8n workflow that writes a first cold email and one follow-up for every lead you send it. You POST a list of leads to a
webhook address. For each lead, a model writes a short subject line, a personal email and a follow-up for three days later.
The workflow saves them in a Google Sheet for you to review. It sends nothing.

What is inside

• workflow.json: the workflow as n8n exports it, seven nodes. One is a sticky note. It holds no credential.
• README.md: these notes.

What we changed from the original copy

• The AI step was an n8n-nodes-base.openAi node at version 1.8. That node type only exists at versions 1 and 1.1, and
  its options differ. AI: Write Cold Email is now n8n's OpenAI node, which has version 1.8, with its Message a Model
  operation and JSON output turned on.
• The model was gpt-4o-mini, which OpenAI shuts down on 14 January 2027. It now asks for gpt-6-luna, the replacement
  OpenAI names. The temperature of 0.7 and the limit of 800 tokens are gone: the model's page does not list a temperature
  setting, and OpenAI counts a reasoning model's thinking against the output limit.
• Merge Lead + Email used the first answer only. A request with several leads would have saved one row. It now pairs
  every answer with its lead.
• Parse Lead List called require('crypto'), which n8n's code runner refuses unless the instance allows it. The
  requestId is now the execution id and the lead's position.
• The sender name, company, value proposition and calendar link could come from four environment variables. Only a
  self-hosted n8n can set them, so the code now reads the request and falls back to plain defaults that you can edit.
• The sheet id came from an environment variable. It is now the placeholder <YOUR_SHEET_ID>. The sheet node also lacked
  its column list, which current n8n requires, so it now lists its nine columns. The tab is called Lead emails.
• The webhook path was the author's product name. It is now lead-emails, and the webhook id is gone.

Before you start

• An OpenAI API key.
• A Google account for a Google Sheets OAuth2 credential, and a Google Sheet with a tab called Lead emails. Row 1 holds
  these names: Name, Email, Title, Company, Subject Line, Email Body, Follow-up, Generated At and Sent.
  If one is missing, the sheet step stops.

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 AI: Write Cold Email.
3. Add a Google Sheets OAuth2 credential and select it in Save to Google Sheets. Replace <YOUR_SHEET_ID> with your
   sheet's id, the part of its address between /d/ and /edit.
4. In Parse Lead List, change the defaults for the sender name, the company, the value proposition and the calendar
   link, or send them in each request.
5. Publish the workflow. Send a POST request to <YOUR_N8N_HOST>/webhook/lead-emails with a JSON body like the one in
   the example. While you test, use the test address <YOUR_N8N_HOST>/webhook-test/lead-emails after choosing Execute
   workflow.

How a run goes

1. Receive Lead Data takes the POST and waits for the workflow to answer.
2. Parse Lead List reads leads from the body, or treats the body as one lead. A lead needs at least a name.
   It adds the sender details and stops with a clear message when the list is empty.
3. AI: Write Cold Email asks the model for JSON: a subject line of 3 to 7 words, an email body of at most 80 words, and
   a follow-up of 30 to 50 words for three days later.
4. Merge Lead + Email joins each answer to its lead.
5. Save to Google Sheets appends a row for each lead, with Sent set to No, and Respond: Success answers the request
   with the lead, the company and the subject line.

Change the result

• The voice and the rules of the email: the system message in AI: Write Cold Email.
• The sender details: the defaults in Parse Lead List, or sender_name, sender_company, value_prop and
  calendar_link in the request.
• The extra facts the model can use: industry, company_description, linkedin_url and notes on each lead.
• The columns: the mapping in Save to Google Sheets and row 1 of your tab.

Good to know

• The webhook has no authentication, as in the original copy. Anyone who knows the address can spend your OpenAI credit.
  Unpublish the workflow when you do not use it, or set a Header Auth credential on the Receive Lead Data node and send
  the header with each request.
• Nothing is sent. The sheet only holds drafts, and the Sent column starts as No.
• Each lead is one model call, and the request waits until all of them are done. The sticky note suggests lists of 5 to 20.
• The follow-up's subject is written by the model but not saved. The sheet keeps only the follow-up's text.
• The model may state things about a lead that are not true. Read each email before you send it, and follow the rules for
  cold email in the countries where your leads live.
• Mindket did not run this workflow against OpenAI or Google Sheets. These notes follow n8n's published node code and
  OpenAI's pages, read on 12 November 2026, and the two Code steps were run on sample leads on a local computer.