How to Set Up an AI Workflow to Extract Data from PDF Faxes and Send It to Google Sheets or a Database

Photo by ZHENYU LUO on Unsplash

AI gives new ways to perform daily tasks. Especially those jobs that require more time to extract data and copy it to other sources. Fax is still a medium of communication in healthcare, finance, logistics, or law. AI provides an intelligent way to extract key data from PDF faxes and send it straight to Google Sheets or a database.

Why Automate Fax Data Extraction?

• Save time: Manual data entry is tedious, slow, and expensive.
• Reduce errors: Automated systems can eliminate copy-paste mistakes.
• Scale effortlessly: Handle 10 faxes or 10,000 — same pipeline.
• Integrate with your stack: Data goes directly into the tools you already use — like Google Sheets, PostgreSQL, or Airtable.

What You’ll Build

• Receives a fax as a PDF (via email or a fax-to-digital service).
• Extracts text from the PDF using Optical Character Recognition (OCR).
• Parses the relevant fields using AI or smart rules.
• Sends the extracted data to Google Sheets or a database.

Receive the fax as a PDF

First, make sure your faxes are coming in digitally. You can use any fax-to-email service like

• eFax
• Fax Plus
• HelloFax
• MyFax

Extract Text Using OCR

I remember the time when I tried to develop an OCR system to read government documents, and it took 1 year to develop a system. But AI has the power to build a system in less than a month.

• Tesseract OCR
• Google Cloud Vision OCR
• AWS Textract
• Adobe PDF Services API

AI-Powered Extraction

• ChatGPT or GPT-4 API for structured parsing
• Hugging Face NER models for named entity recognition

Prompt example for GPT:

Extract the date, invoice number, and total amount from the following text. Return as JSON.

{
  "date": "03/12/2024",
  "invoice_number": "INV-2394",
  "total_amount": "$1,250.00"
}

Push Data to Google Sheets or a Database

Once you have clean, structured data, it’s time to send it somewhere useful.

Google Sheets (via API)

Use gspread in Python:

import gspread
from oauth2client.service_account import ServiceAccountCredentials
#Your function to send the data
def write_to_sheet(data):
    
    Googlescope = ["https://spreadsheets.google.com/feeds", "https://www.googleapis.com/auth/drive"]
    #Goolge URLS for data
    creds = ServiceAccountCredentials.from_json_keyfile_name('creds.json', Googlescope)
    client = gspread.authorize(creds)
    sheet = client.open("Fax Data").sheet1
    sheet.append_row([data["date"], data["invoice_number"], data["total_amount"]])

PostgreSQL or MySQL

import psycopg2

def write_to_db(data):
    conn = psycopg2.connect("dbname=faxdata user=postgres password=secret")
    cur = conn.cursor()
    cur.execute(
        "INSERT INTO invoices (date, invoice_number, total_amount) VALUES (%s, %s, %s)",
        (data["date"], data["invoice_number"], data["total_amount"])
    )
    conn.commit()
    cur.close()
    conn.close()

Use Zapier or Make.com for No-Code Automation

you want to skip all the code and move faster, services like Zapier, Make.com, or n8n allow you to:

• Watch a folder for new PDFs
• Run OCR and parsing via cloud APIs
• Insert results into Google Sheets, Airtable, or a database — all with visual workflows

Combine tools like:

• Google Drive
• PDF.co (for OCR)
• ChatGPT via webhook (for field extraction)
• Google Sheets or Airtable

In the above, AI-powered OCR and simple automation tools free yourself from the drudgery of manual data entry. Next step? Set it up. Then enjoy your coffee while the bots do the boring stuff.