Copyright © 2026 Authors retain the copyright of this article. This article 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.

@article{197121,
        author = {Dr B. Suri Babu and K. Chandra Bhanu and Y.B.S. Phaneendra and G.V.S.S. Vara Prasad and M. Abhi and Dr Yalla Venkat},
        title = {AI-Driven Consumer Complaint Intelligence System by using Hybrid Machine Learning and Generative AI with Enterprise N-Tier Architecture},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {6177-6183},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=197121},
        abstract = {Organizations across banking, e-commerce, healthcare, and telecommunications receive thousands of consumer complaints daily through digital channels. Existing complaint management systems fail on three critical dimensions: they cannot prioritise emergencies dynamically, they offer no transparency into automated routing decisions, and they lack the architectural integrity required for enterprise-scale deployment. This paper presents an AI-Driven Consumer Complaint Intelligence System that addresses these gaps through a hybrid three-paradigm intelligence strategy. A Predictive Machine Learning pipeline employing TF-IDF vectorisation and Logistic Regression classifies unstructured complaint text into six product categories. A Rule-Based Severity Assessment Engine deterministically assigns priority levels (P1 Emergency, P2 Operational, P3 General) using keyword heuristics at near-zero latency. A Generative AI layer powered by Google Gemini produces human-readable explanations for every classification decision and autonomously drafts professional email responses. These intelligence vectors are orchestrated through a Decision Engine Matrix, routing each complaint to Auto-Send, Review Required, or Escalate outcomes. The entire pipeline is deployed within a Five-Layer Enterprise N-Tier Architecture (React SPA, Flask REST API, Hybrid AI Tier, SQL Alchemy ORM, PostgreSQL). On a balanced 1,260-record dataset, the classification model achieved 94% accuracy on a held-out test set of 252 samples, the severity engine produced sub-10 ms detection latency, and the P1 false-negative rate was 0% in the evaluated scope.},
        keywords = {Consumer Complaint Intelligence; Hybrid AI; Machine Learning; Generative AI; Severity Assessment; N-Tier Architecture; Explainable AI; TF-IDF; Logistic Regression},
        month = {April},
        }