Adaptive Assistants for Customized E-Shopping

benefits such as convenience and low prices. But such benefits are partly offset by a new cost—the time and frustration required to sift through information on innumerable vendor sites and to learn how to submit and manage orders. These tasks are difficult because the type, amount, and organization of information differ from vendor to vendor. Complicating matters, customers are unaware of changes in pricing, availability, and so on unless they visit the same sites frequently, removing a potential advantage of online stores over traditional ones. Can intelligent agents improve the e-shopping experience? Currently, online shoppers can adopt a number of strategies when looking for a product. The most straightforward approach is to visit various vendor sites; for each site, the shopper browses or searches for a particular product. This simple approach has several drawbacks. First, because no single site caters to all shopping needs, the user’s search time increases for each new product category. Second, getting acquainted with individual nonstandard vendor interfaces slows browsing and hinders impulse shopping. Third, this approach likely favors only the largest vendors (owing to name-branding), which reduces the market’s efficiency by providing fewer competitive choices to consumers. Several services let shoppers sign up to receive price alerts that notify them when a product’s price changes or falls below a specified amount. Some of these services require shoppers to fill out lengthy surveys, yet most of the sites offer little or no customization. This shopping approach also weakens user privacy. Another solution involves compilation of voluntary user ratings and reviews of vendors and products.1 Such recommendation systems might reduce the marketplace’s size and introduce bias, because obtaining a sufficient number of ratings for every vendor and controlling the sources’ reliability are difficult. Another alternative further automates and generalizes the search.2,3 As early as 1995, researchers proposed comparison-shopping agents (also known as shopping bots) as a solution for finding products under the best terms (price was typically the most important feature) among vendor sites. A shopping agent queries multiple sites on behalf of a shopper to gather pricing and other information on products and services. Most comparison-shopping agents, however, present a marketplace that is biased in favor of the vendor sites that collaborate with (pay fees to) the shopping agent. In addition, a shopper has only a limited number of vendor sites to choose from, and often the participating sites do not offer the best prices. (For more on shopping agents, see the related sidebar.) Intelligent customization techniques can greatly improve the accessibility and consumer benefits of e-shopping by creating a personalized (and thus more efficient) marketplace. We designed and built IntelliShopper, a new type of shopping agent that can provide customers with adaptive and customized shopping assistance. IntelliShopper learns users’personal preferences and autonomously shops on their behalf while protecting their privacy.