PartsID: A Dialogue-Based System for Identifying Parts for Medical Systems

PartslD: A Dialogue-Based System for Identifying Parts for Medical Systems Amit BAGGA, Tomek STRZALKOWSKI, and G. Bowden WISE Information Technology Laboratory GE Corporate Research and Development 1 Research Circle Niskayuna, USA, NY 12309 { bagga, strzalkowski, wisegb } @crd.ge.com Abstract This paper describes a system that provides customer service by allowing users to retrieve identification numbers of parts for medical systems using spoken natural language dialogue. The paper also presents an evaluation of the system which shows that the system successfully retrieves the identification numbers of approximately 80% of the parts. Introduction Currently people deal with customer service centers either over the phone or on the world wide web on a regular basis. These service centers support a wide variety of tasks including checking the balance of a bank or a credit card account, transferring money from one account to another, buying airline tickets, and filing one's income tax returns. Most of these customer service centers use interactive voice response (IVR) systems on the front-end for determining the user's need by providing a list of options that the user can choose from, and then routing the call appropriately. The IVRs also gather essential information like the user's bank account number, social security number, etc. For back-end support, the customer service centers use either specialized computer systems (example: a system that retrieves the account balance from a database), or, as in most cases, human operators. However, the IVR systems are unwieldy to use. Often a user's needs are not covered by the options provided by the system forcing the user to hit 0 to transfer to a human operator. In addition, frequent users often memorize the sequence of options that will get them the desired information. Therefore, any change in the options greatly inconveniences these users. Moreover, there are users that always hit 0 to speak to a live operator because they prefer to deal with a human instead of a machine. Finally, as customer service providers continue to rapidly add functionality to their IVR systems, the size and complexity of these systems continues to grow proportionally. In some popular systems like the IVR system that provides customer service for the Internal Revenue Service (IRS), the user is initially bombarded with 10 different options with each option leading to sub-menus offering a further 3- 5 options, and so on. The total number of nodes in the tree corresponding to the IRS' IVR system is quite large (approximately 100) making it extremely complex to use. Some customer service providers have started to take advantage of the recent advances in speech recognition technology. Therefore, some of the IVR systems now allow users to say the option number (1, 2, 3 ..... etc.) instead of pressing the corresponding button. In addition, some providers have taken this a step further by allowing users to say a keyword or a phrase from a list of keywords and/or phrases. For example, AT&T, the long distance company, provides their users the following options: "Please say information for information …

Content truncated to 500 words