<aside> <img src="/icons/command-line_gray.svg" alt="/icons/command-line_gray.svg" width="40px" />
Goal: Automate the extraction of key business intelligence from unstructured text sources, such as meeting transcripts, to save time and improve data accessibility for teams.
How It Works (Zapier Flow Breakdown):
🔹 Step 1 – Trigger: Google Docs When a new doc is added to a specific folder ("BOTS"), the automation starts.
🔹 Step 2 – Action: Google AI Studio (Gemini)
Process the document content using the Gemini API to identify and extract key business information, such as client names, project requirements, and action items Sends the document's content to Gemini with this prompt: “You are an expert assistant tasked with extracting key business information from a meeting transcript…” Gemini returns a clean JSON object with:
{ "business_official_name": “name”, "public_email": “email”, ... }
🔹 Step 3 – Action: Code by Zapier (Python)
Python script:
Parses and validates JSON
Removes nulls, whitespace, errors
Ensures proper formatting before database entry
🔹 Step 4 – Action: Google Sheets Inserts the cleaned data into a spreadsheet under predefined columns.
Tools Used: Google Docs, Gemini AI (Google AI Studio), Python via Code by Zapier, Google Sheets and Zapier.
Python Script:
Code for the JSON cleaning step:
🔗 https://github.com/AlvLeoAI/zapier/blob/main/ai_business_extractor.py
Result:
<aside> <img src="/icons/link_gray.svg" alt="/icons/link_gray.svg" width="40px" />
https://zapier.com/shared/8b76bdf926bdee61f8c6eac2c223e5e8ee5b5ab4
</aside>