Introducing Chatbots in Libraries
Michele L. McNeal, David Newyear · Library Technology Reports · 2013
Abstract Chapter 1 of Library Technology Reports (vol. 49, no. 8), Streamlining Information Services Using Chatbots, presents a brief history of chatbots, computer programs that use natural language to interact with users. They have existed for nearly fifty years and have been used in libraries since mid-2000s; from ELIZA (1966) to Pixel (2010) are introduced. ********** As many libraries continue to see reductions in funding, we are increasingly seeing technology as a way to make up for budget shortfalls. In circulation context, online patron account management, self-registration, and self-serve checkout stations are examples of this trend. Since requests for basic library information (including locations, hours, and policies) and for specific materials or resources predominate among chat and IM inquiries of libraries, (1) chatbots or offer a self-service option for our online customers in context of information services. Such virtual agents are becoming a familiar feature on many websites. These user-friendly implementations of artificial intelligence have enjoyed remarkable success in corporate and government sectors and are projected to continue to grow in popularity. Indeed, a VirtuOz/CCM Benchmark Group study projects a 400% increase in virtual agents between 2011 and 2014. (2) Chatbots are able to respond to a remarkable variety of customer inquiries with correct information specifically tailored to customer's needs through use of natural language processing (NLP). Warschauer and Healy define NLP as the process of a computer extracting meaningful information from natural language input and/or producing natural language output. (3) The process of searching databases or catalogs usually requires user to compose a search for information needed, conforming to structures and language defined by target data source. A chatbot using NLP, on other hand, allows users to pose a question as they would to another human being. The responsibility of locating needed information shifts from user to programmer of chatbot. The chatbot designer creates a structure that leads user through a question-and-answer dialogue to discover information needed and to provide it. This process can also address problems created by library terminology or jargon with which user may not be familiar. In addition, regular review of chatbot's conversation logs allows designer to monitor types of questions and terminology used to pose them and to update responses provided by chatbot and language it recognizes. This is why chatbot can be particularly convenient and helpful to those patrons who are least familiar with library and its services. * Chatbots are also consistently patient and polite and remain unruffled by rude customers, high traffic, or repeated requests for same information. In a discussion of chatbots, Christansen suggests that chatbots: ** were selected more frequently than other forms of digital reference ** made asking questions easier (by providing a natural language interface) ** provided instant responses ** were anonymous (which encouraged shy users or those who thought their questions might be stupid) ** provided a marketing tool for reference services (4) Christiansen's final observation is particularly noteworthy. If we enhance our hot with links to our various resources (including our staff directory, our catalog, and our online databases and reference tools), we can introduce our customers to our whole gamut of useful resources. In addition, offer ability to personalize user experience, welcoming patrons and putting them at ease. This is highlighted in a SitePal case study, which documented a 40 percent increase in sales at a company that incorporated a chatbot to welcome users to its website and guide them through its broad inventory of products. …