AI Prospecting Automation with n8n
An end-to-end workflow that reads documents with OCR, writes personalised emails with an LLM, sends them through Gmail and tracks everything in Google Sheets.
Overview
A no-code / low-code automation I built to run my own internship and job prospecting: it extracts information from documents, generates a tailored message for each contact with an LLM, sends it with the Gmail API and logs every step in Google Sheets.
The problem
Prospecting by hand meant copying contact details from documents, rewriting similar emails dozens of times and keeping track of who had been contacted — hours of repetitive work with frequent mistakes.
What I built
I built an n8n workflow: OCR extracts text from documents, an LLM (DeepSeek) structures the data and drafts a personalised email, the Gmail API sends it, and a Google Sheet stores each contact, message and status so follow-ups are easy. Prompts are versioned and every step can be re-run.
Results
- Prospecting emails generated and sent automatically, each one personalised
- One tracking sheet with the status of every contact
- The same pattern reusable for lead generation, invoices or support emails
Highlights
- OCR → LLM → Gmail → Google Sheets pipeline
- Prompt engineering for structured extraction and email writing
- Error handling and re-runnable steps in n8n
