Technical Architecture:

1. Data Collection: Fetches earnings report PDFs from links stored in Google Sheets
2. Document Processing: Downloads and parses PDFs into semantically meaningful chunks
3. Vector Embedding: Converts text chunks into embeddings using Google's text-embedding-004 model
4. Semantic Database: Stores embeddings in Pinecone for intelligent retrieval based on meaning
5. AI Analysis: Utilizes GPT-4o-mini and Gemini AI to interpret data and generate insights
6. Report Generation: Automatically compiles findings into a structured Google Doc

This seamless integration of document processing, vector search and multiple AI models creates a system that can understand financial context, identify trends, and generate meaningful analysis without constant human supervision.