This paper describes a web-based dialog system - Natural Language Sales Assistant (NLSA) - that helps users find relevant information about products and services in ecommerce sites. The system leverages technologies in natural language processing and human computer interaction to create a faster and more intuitive way of interacting with websites. By combining traditional AI rulebased technology with taxonomy mapping, the system is able to accommodate both customer and business requirements. Our user studies have demonstrated that, in the context of e-commerce, users preferred the natural language enabled navigation over menu-driven navigation (79% to 21% users). In addition, compared to a menu driven system, the average number of clicks used in the natural language system was reduced by 63.2% and the average time was reduced by 33.3%. The NLSA system is currently deployed by IBM as a live pilot and we are collecting real user feedback. We believe that conversational interfaces like that of NLSA offer the ultimate personalization and can greatly enhance the user experience for websites.