In modern digital platforms, optimizing user experience (UX) is crucial for user engagement and business success. Traditional A/B testing methods are widely used but can be time-consuming, require a lot of traffic, and struggle to adjust dynamically. To tackle these issues, we propose an AI-enhanced A/B testing framework that combines machine learning models and adaptive decision-making algorithms to optimize UX more efficiently. Our approach uses predictive modeling to estimate design performance with smaller datasets, which shortens the duration of experiments. We also include a multi-armed bandit strategy that reallocates user traffic to better-performing design variants in real time, reducing the costs of poor- performing options. The system incorporates detailed behavioral analytics, like cursor movements, scroll depth, hesitation patterns, and engagement metrics. This provides deeper insights into user interactions beyond standard conversion rates. This AI-driven approach speeds up decision-making and lowers experimental overhead, ensuring ongoing adjustment to changing user behavior. By connecting UX research with AI-driven analytics, our framework gives organizations a smart, scalable way to improve UX iteratively.
S. Dhanalakshmi (2025) studied this question.