Food Complaint Resolution

Created by team DevStat Fusion on January 22, 2024

The food complaint resolution application is an AI-based customer complaint system tailored to the food delivery industry. While many food delivery companies already have a customer complaint system, it typically involves human intervention to address customer concerns. For instance, if a customer orders pita bread and receives expired bread, filing a complaint triggers a human review process. The human evaluator assesses the complaint using metadata such as images of the bread and the complaint description. Following this evaluation, the customer may be reimbursed if the human determines that the delivered pita bread has indeed expired. However, this manual process is time-consuming and dependent on the availability of human personnel. By utilizing the food complaint resolution application, a food delivery company can significantly reduce processing time and human effort in handling complaints. The application can swiftly process customer complaints in a matter of seconds, streamlining the resolution process. We developed this application with ❤️, incorporating state-of-the-art AI models such as Food Item recognition, GPT-4 Turbo, and GPT-4 vision to automate complaint resolution. The user-friendly interface, created using the Streamlit library, ensures a seamless experience for users. The system mimics human evaluation by processing complaints based on metadata like complaint descriptions and images. After analyzing the information provided by the customer, our system adheres to company policies to provide cash back to eligible customers. If a customer remains dissatisfied with the refund or after reaching the maximum attempts, the system redirects the complaint to human intervention.

Category tags:

"Solid concept and idea, and I like how unique it is. I can see this being directly deployed in many food delivery platform businesses. So I hope you guys can keep working on the product. Best of luck."


Shebagi Mitra

Technical Mentor

"In the food market, it is impactful. A lot of people ordered food and found damages. Helpful for both sides. Good work. "


Muhammad Mubashir Hassan


"Great work, interesting idea and simple application. However, it seems like your custom model food-image-recognition seems to be not generating results using external PAT keys. But overall, good work, by results can be improved using RAG implementation. Keep it up. "


Muhammad Inaamullah

Machine Learning Engineer