IntelliBargainBot: AI-Driven Price Negotiation System for E-Commerce Platforms

Abstract

The rapid growth of e-commerce has significantly improved online shopping experiences. However, the traditional bargaining experience is largely absent in most e-commerce platforms due to fixed pricing mechanisms, limiting customer interaction and pricing flexibility. This paper presents IntelliBargainBot, an AI-driven price negotiation system that enables personalized bargaining through the integration of machine learning, natural language processing (NLP), and automated negotiation techniques. Customer transaction data is transformed into RFM and behavioral features, which are used by the K-Means clustering algorithm to segment customers based on purchasing patterns. The identified customer segment guides negotiation strategies, including discount limits, concession behavior, negotiation rounds, and minimum acceptable prices. Product demand and purchase quantity are also considered to personalize negotiation outcomes. A hybrid NLP module combines rule-based keyword matching with transformer-based intent classification to identify customer negotiation intents. Based on the detected intent and customer profile, a polynomial concession model generates adaptive counter-offers, while a template-based response generation module delivers clear and consistent negotiation responses. Experimental results demonstrate that the proposed framework effectively integrates customer segmentation, intelligent intent detection, and adaptive negotiation strategies to improve customer engagement while maintaining seller profitability in e-commerce platforms.

Introduction

This paper proposes IntelliBargainBot, an AI-powered price negotiation system for e-commerce platforms that introduces personalized bargaining into online shopping. Unlike conventional e-commerce platforms that rely on fixed pricing, the proposed system enables customers to negotiate product prices through an intelligent conversational chatbot. By integrating machine learning, natural language processing (NLP), and automated negotiation techniques, IntelliBargainBot analyzes customer behavior and negotiation intent to generate personalized counter-offers while ensuring business profitability.

The primary motivation behind the system is that existing online shopping platforms generally ignore factors such as customer loyalty, purchasing behavior, negotiation history, and product demand when determining prices. This limits customer engagement and eliminates the flexible bargaining experience commonly found in traditional marketplaces. IntelliBargainBot addresses this limitation by supporting adaptive price negotiation based on customer characteristics and business constraints.

The main objectives of the proposed system are to analyze customer purchasing behavior using Recency, Frequency, and Monetary (RFM) features, segment customers through the K-Means clustering algorithm, detect customer negotiation intent using a hybrid NLP approach, generate personalized counter-offers based on customer segment, product demand, and purchase quantity, and improve customer engagement without compromising seller profitability.

The literature review highlights previous work on AI-driven negotiation chatbots that combine machine learning, NLP, dialogue management, and negotiation strategies. Existing systems support conversational bargaining, text and voice interactions, intent detection, and rule-based negotiation. However, the proposed IntelliBargainBot distinguishes itself by integrating customer segmentation, hybrid intent detection, adaptive negotiation policies, and dynamic concession strategies into a unified bargaining framework.

The proposed architecture consists of a React.js frontend, a Flask backend, and Python-based machine learning and NLP modules. Customers browse products through the e-commerce website and initiate negotiations using the BargainBot interface. The backend retrieves customer and product information, predicts the customer's segment using the trained K-Means model, detects negotiation intent through the hybrid NLP module, generates negotiation policies using predefined business rules, computes adaptive counter-offers through a polynomial concession model, and finally returns personalized responses using template-based response generation.

Customer segmentation is developed using an online retail transaction dataset. After data collection and preprocessing, customer behavior is represented using RFM metrics along with additional behavioral features. These features are used to train a K-Means clustering model that classifies customers into three meaningful segments: New, Regular, and Loyal. During live negotiations, the trained model predicts the customer's segment, enabling the system to personalize negotiation strategies.

The Hybrid NLP-Based Intent Detection module combines rule-based keyword matching with a transformer-based intent classification model. Common negotiation intents such as offer submission, acceptance, and rejection are first identified using predefined keywords. If keyword matching is insufficient, a transformer model analyzes the semantic meaning of the customer's message. Regular expression techniques are also used to extract the customer's proposed price from negotiation messages before passing the information to the negotiation engine.

The Rule-Based Negotiation Policy Engine generates personalized negotiation strategies according to customer segment, product demand, and purchase quantity. It determines parameters such as maximum discount, minimum acceptable price, concession rate, and the maximum number of negotiation rounds. Bulk purchase discounts are also incorporated to support quantity-based negotiations while preserving profitability.

The Polynomial Concession Model computes adaptive counter-offers by gradually adjusting prices across negotiation rounds. The concession behavior is controlled by predefined concession parameters and customer offers while ensuring that prices never fall below the minimum acceptable threshold. This approach simulates realistic bargaining behavior and provides more personalized negotiations.

The Template-Based Response Generation module produces consistent conversational responses, including counter-offers, offer acceptance, rejection, clarification requests, final offers, and bulk purchase responses. Using predefined templates ensures efficient real-time interaction while maintaining conversational quality.

The system's performance is evaluated using clustering metrics such as the Elbow Method and Silhouette Score. Experimental analysis shows that although two clusters achieve the highest Silhouette Score, three clusters (K = 3) are selected because they provide meaningful customer categories—New, Regular, and Loyal—that better support personalized negotiation strategies. Additional analysis includes feature correlation heatmaps and Principal Component Analysis (PCA) visualizations, demonstrating the effectiveness of the K-Means clustering model in separating customer segments.

Conclusion

This paper presented IntelliBargainBot, an intelligent price negotiation system that enables personalized bargaining in e-commerce platforms. Experimental results demonstrated that the proposed framework effectively supports customer-specific negotiation strategies, resulting in flexible and personalized pricing while maintaining predefined business constraints. The integration of customer segmentation and intelligent negotiation mechanisms improves customer engagement, enhances the online shopping experience, and supports fair pricing decisions. Overall, the proposed system provides an effective solution for incorporating automated price negotiation into modern e-commerce platforms while balancing customer satisfaction and business profitability.

References

[1] B. Naseeba, A. S. Sathwik, J. C. Kiran, and N. P. Challa, “An Intelligent Chatbot Haggling with Ensemble ML Model,” in Proc. 3rd Int. Conf. Pervasive Comput. Social Netw. (ICPCSN), IEEE, 2023, doi: 10.1109/ICPCSN58827.2023.00034. [2] P. D. Sree, M. R. Kokkiligadda, J. Teja, and Y. Sandeep, “Product Negotiation in E-Commerce Website using Chatbot,” in 2023 7th International Conference on Computing Methodologies and Communication (ICCMC), Erode, India, Feb. 2023, doi: 10.1109/ICCMC56507.2023.10083745. [3] B. Usha Sri, P. Aakanksha, G. Tirumala Dev, K. Yasoda Krishna, and I. Gayathri, “Price Negotiation Chatbots on E-Commerce Website Using Machine Learning,” International Journal of Creative Research Thoughts (IJCRT), vol. 12, no. 5, pp. d481–d486, May 2024. [4] Z. Zhao, M. Theune, S. Srivastava, and D. Braun, “Exploring Lexical Alignment in a Price Bargain Chatbot,” in Proc. 6th ACM Conference on Conversational User Interfaces (CUI ’ 24), Luxembourg, July 8-10 2024, ACM, 7 pp., doi: 10.1145/3640794.3665576. [5] S. Pappala, S. Rakesh, V. Gangaiah, G. Poojitha, and T. Madhu, “Smart Bargain Bot: A Text and Voice-Based Price Negotiation System for E-Commerce Platforms,” International Journal on Science and Technology (IJSAT), vol. 16, no. 2, pp. 1–11, Apr.–Jun. 2025. [6] R. El Youbi, F. Messaoudi, and M. Loukili, “Machine Learning-driven Dynamic Pricing Strategies in E-Commerce,” in Proc. 2023 14th International Conference on Information and Communication Systems (ICICS), IEEE, 2023, doi: 10.1109/ICICS60529.2023.10330541. [7] D. Ramachandran, A. Keshari, and M. K. Tiwari, “Contract price negotiation using an AI-based chatbot,” in Applications of Emerging Technologies and AI/ML Algorithms, Springer, Singapore, 2023, pp. 303–310. doi: 10.1007/978-981-99-1019-9_27. [8] O. N. Akande, H. B. Akande, E. O. Asani, and B. T. Dautare, “Customer Segmentation through RFM Analysis and K-means Clustering: Leveraging Data-Driven Insights for Effective Marketing Strategy,” in Proc. 2024 International Conference on Science, Engineering and Business for Driving Sustainable Development Goals (SEB4SDG), IEEE, 2024, doi: 10.1109/SEB4SDG60871.2024.10630052.

Copyright

Copyright © 2026 Erukulla Pravalika, Dr. V. Uma Rani . 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.