Automations and Integrations Projects

2 Extraction data from PDF with AI-powered automated process

Read (OCR scan) a PDF document containing shipping information, where an automated AI process using OpenAI LLM model, structures the required data into JSON format, which is then exported to Google Sheets for analysis and cross-referencing with the business's bank transactions. This enables detailed control over shipments delivered to customers.

Application Architecture

• Google Drive: stores PDF documents and allows downloading documents to carry out the process
• PDF.CO: platform that provides API for scan and extract data
• OpenAI: Model LLM (GPT-4.o Mini 2024) — structures data in JSON format
• Google Sheet: to export the final scanned data from PDF

Final Implementation

1. PDF documents containing shipping information are uploaded to a Google Drive folder, which stores the files and allows them to be downloaded for processing.
2. The required data is scanned from the PDF using the PDF.co platform API and extracted as plain text.
3. The OpenAI LLM model (GPT-4.o Mini 2024) is used to precisely extract the necessary data and structure it in JSON format, properly labeling each value.

Finally, the structured JSON data is mapped and exported to Google Sheets in order to cross-reference it with the business's bank transactions and maintain control over shipments delivered to customers.

Graphical Technical Design

1 Automations process and IA Generative Tools — MAKE & N8N

Hand on-lab: Using tools like MAKE and N8N will allow me to create and deploy workflows where repetitive tasks are automated and improve the productivity of business processes through automation and integrations with AI. N8N service was implemented under AWS cloud architecture for my own professional work in automation and AI Agents.

Graphical Technical Design