Personalizing E-commerce applications with on-line heuristic decision making

This paper describes new technology based on on-line decision support for providing personalized customer treatments in web-based storefronts and information sites. The central improvement over existing systems is a new paradigm for specifying decisions, based on a language that incorporates owchart constructs, rules-based constructs, and a variety of specialized constructs to facilitate reasoning based on heuristics and partial information. Reports about decisions made by a program in this language have structure that is conceptually close to the structure of that program. This makes it easy for business analysts and managers to tune the programs to enhance business performance. To illustrate the bene ts of our approach this paper describes the May-I-Help-You (MIHU) prototype system, that monitors a customer's progress through a web storefront, and may choose to proactively intervene in order to help close a sale. The intervention might o er a discount or promotion, or give the customer a \May I Help You" window, that o ers an opportunity to have text chat, voice chat, and/or escorted browsing with a Customer Service Representative (CSR). In MIHU, the decision about whether to o er live assistance is fully automated, taking into account not only the business value of a given customer interaction, but also the current availability of CSRs to help realize this opportunity.