A high-performance AI workflow designed to audit massive (200+ page) technical and legal documents with 100% citation accuracy.
Traditional RAG (Retrieval-Augmented Generation) often suffers from "Lost in the Middle" syndrome, where AI misses details in the center of long documents. This engine utilizes Gemini 1.5 Pro's 2-million-token context window to maintain full visibility across the entire dataset.
- Engine: Google AI Studio (Gemini 3.1 Pro)
- Logic: Zero-Shot System Instruction with Temperature 0.0
- Input: [e.g., EU AI Act / AWS Security Framework]
- Automated Risk Matrix: A table of all financial penalties and compliance deadlines.
- Citation Verification: Every claim includes a specific Page Number and Section Title.
- Cross-Document Synthesis: Linked Section 4 (Data Privacy) directly to Section 12 (Enforcement).
Developed as part of an AI Productivity Portfolio for High-Ticket Consulting.
URL to the prompt and model response for reference :- https://aistudio.google.com/app/prompts?state=%7B%22ids%22:%5B%221N8nPgWj13UhAHCYG-2GV-8zxSfhQPa4y%22%5D,%22action%22:%22open%22,%22userId%22:%22116080054008612041388%22,%22resourceKeys%22:%7B%7D%7D&usp=sharing