Cloud-Based Intelligent Complaint Registration and Analysis System using GenAI and Semantic Triage

Abstract

Complaint management systems are very important for handling user complaints in an efficient way in todays digital world. Traditional systems mainly rely on people to sort and route complaints, which can cause delays and mistakes. Recent improvements in computer language understanding and smart automation have made it possible to make these systems more intelligent.This paper talks about a complaint registration and analysis system that uses cloud computing and a mix of artificial intelligence tools. The system uses Amazons API Gateway and Lambda for handling lots of data Amazon S3 for storing data and AWS Glue with Apache Spark for processing data. A simple classifier is used to sort complaints at first while Googles AI models summarize complaints suggest who should handle them and figure out what they are really about.The system also has a feature that analyzes complaints to understand what users really mean and sends them to the departments with explanations and confidence levels. There\'s a chatbot, for users and administrators to interact with. The proposed system shows accuracy faster response times and better decision-making compared to old complaint management systems.

Introduction

The proposed system uses NLP, machine learning, and generative AI to automatically process, analyze, and route complaints. Users submit complaints through a web interface, which are stored in AWS S3 and processed through a serverless pipeline using API Gateway, Lambda, Glue, and Spark. The system first cleans and preprocesses text, then classifies it using a Naive Bayes model, followed by deeper analysis using a GenAI model (e.g., Gemini) for summarization, validation, and contextual understanding. A semantic triage module then assigns each complaint to the correct department based on meaning rather than simple keyword rules.

A key feature is the hybrid AI approach, combining fast traditional machine learning with advanced generative AI reasoning to improve accuracy, context awareness, and decision-making. The system also includes a chatbot interface for user interaction and transparency.

Results show strong performance: around 85–90% classification accuracy, about 92% accuracy in routing complaints correctly, and an average response time under 2 seconds. The serverless AWS architecture ensures scalability and efficient handling of high workloads without performance degradation.

Conclusion

This paper talks about a complaint management system that uses the artificial intelligence to analyze and route complaints automatically.It uses a mix of two types of AI, NLP and GenAI to make the system work better.The way it uses AI and a special triage mechanism makes it more efficient and accurate. The system shows how cloud computing and AI can be used together to make applications that\'re both smart and can handle a lot of work.It uses cloud computing and artificial intelligence to make complaint management system more scalable and intelligent.The system can handle complaints.It helps to resolve issues.

References

[1] R. Harish Selva Pandi, M. Sivabalan, M. Dhanush, and G. Praveen Kumar, “Smart Complaint and Feedback System Using NLP,” International Research Journal on Advanced Engineering Hub, vol. 4, no. 2, pp. 547–554, 2026. [2] Qurat-ul-Ain, A. Shaukat, and U. Saif, “NLP Based Model for Classification of Complaints: Autonomous and Intelligent System,” in Proceedings of the IEEE Conference on Digital Futures, 2022. [3] C. K. Mankala et al., “Sustainable Real-Time NLP with Serverless Parallel Architecture,” Information Journal, 2025. [4] S. O. Amelia and F. Al-Fulan, “Designing Serverless Data Lakes with Amazon S3, AWS Glue, and AWS Lambda,” ResearchGate, 2025. [5] D. Vuppu, “Serverless ETL: Leveraging AWS Glue and PySpark for Efficient Data Processing,” ResearchGate, 2025. [6] L. Thompson and R. White, “Exploring the Scalability of Serverless Architectures with AWS Lambda and S3,” Journal of Cloud Computing Research, 2018. [7] M. Garcia and S. Patel, “Analyzing Cost and Performance Trade-offs in Serverless Data Processing Pipelines,” IEEE Transactions on Cloud Computing, 2019. [8] R. Kumar and V. Singh, “Leveraging AWS Glue for Large-Scale ETL Operations in Modern Data Lakes,” in Proceedings of the International Conference on Data Engineering, 2020. [9] T. Nguyen, J. Lee, and S. Park, “A Serverless Data Ingestion Framework Using AWS Lambda,” International Journal of Distributed Sensor Networks, 2020. [10] Y. Zhao, A. Thompson, and R. Hernandez, “Evaluating Serverless Data Processing Frameworks on Cloud Platforms,” IEEE Transactions on Cloud Computing, 2019. [11] K. Villamizar et al., “Comparative Study of Monolithic and Microservices Architectures on AWS,” International Journal of Cloud Computing, 2016. [12] Y. Kim and J. Lin, “Serverless Data Analytics with Flint,” arXiv preprint arXiv:1803.06354, 2018. [13] M. Jin and N. Aletras, “Complaint Identification in Social Media with Transformer Networks,” arXiv preprint arXiv:2010.10910, 2020. [14] M. Sharma et al., “Building a Legal Dialogue System: Development Process and Challenges,” arXiv preprint arXiv:2109.00381, 2021. [15] R. Sukthanker et al., “Anaphora and Coreference Resolution: A Review,” arXiv preprint arXiv:1805.11824, 2018.

Copyright

Copyright © 2026 Rajan R, Ravikanth S, Hemajothi , Pradeep Raj . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.