A website is a collection of files and information that covers a particular theme and is managed by an individual or organization. However, present websites are complex and lack intuitive navigation, making it challenging for users to locate specific information efficiently. In addition, existing websites lack interactivity beyond basic input forms and static content, even though users crave engaging experiences that respond dynamically to their actions and preferences, leading to increased user engagement and satisfaction. To achieve this requirement, the aim of this paper is the modeling of an enhanced human-computer interaction website system through transfer learning-based automated voice-controlled approach. The methodology used is an extreme programming approach, while the research method is the data collection of e-commerce website historical information, which is then applied to train a transfer learning algorithm called ResNet to generate a model of the web information system. In addition, the Python text-to-speech library will also be adapted and applied to the developed model before integration into the website to facilitate automated voice control features. The system was implemented with JavaScript and Python programming languages. The results of the ResNet training and testing reported an average of 98.7% accuracy, while the new website, after practical experimentation, showed that the user can actually navigate and find information on the website through voice input.