A simple and straightforward Streamlit app that is powered by a mix of LLM, OCR, and object detection to empower approval teams to manage the overwhelming amount of repair tickets of shipment containers :) Check out the prototype here: https://5dc16b035bc5.ngrok-free.app
- AI-Driven Decision Engine Combines rule-based cost vs. age checks with Cohere Aya for nuanced reasoning and YOLOv8 object detection to analyze damage images in real time
- Interactive Dashboard Track and manage tickets through “Manual Review,” “Additional Data Requested,” “AI Approved,” or “AI Disapproved” queues
- Customizable Settings Define cost thresholds by container age and specify which repair codes require image evidence. Update on the fly via the AI Training & Settings panel.
- Seamless SOP Integration Upload PDFs of your Standard Operating Procedures to automatically extract repair codes needing mandatory visual documentation using Mistral OCR
- In-App Chat & Collaboration Chat with the AI agent for clarifications or loop in your team with manual messages