Dialogs builds web, mobile, and business process applications for enterprise and SMB
The main findings to the tested hypotheses and the corresponding implications for research and practice are summarized in Table 11. (1) Given our need to analyze the utility and efficacy of both SDS dialog strategies with respect to achieving a specific goal (Venable 2006), our evaluation aims to test the rigor of both designs by assessing their functional effectiveness (Venable et al. 2012). Furthermore, we aim to outline the strengths and weaknesses of both dialog strategies by empirically testing the user experience in customer service, thereby reducing design uncertainty and risk. Furthermore, the SDS is capable of telling jokes and engaging in simple small talk; nonetheless, to ensure that the system does not lose task orientation, the prompts always end with the question of the respective process step (cf. Figure 5). If the user deviates too far from the actual task so that the system cannot interpret the statement, an error prompt occurs.
In addition, during the formative evaluation, we involve experts who have contributed their experience with dialog systems in customer service to the design of the user experience. However, given the rather moderate focus on customer service in the first design steps, the transferability and generalizability of our research results may be limited. Further studies that exclusively address domain-specific design requirements and design principles (e.g., based on user stories and user focus groups) should complement our findings. Nevertheless, the results from our summative evaluation (cf. Section 5.2), which we conduct in the customer service domain, clearly demonstrate that the design theory is suitable for satisfying the needs of users from the customer service domain.
Make sure you are transparent with your customers and know what they want from the company.
The positive associations of an efficiently and rationally acting machine should be combined with the a human interlocutor (R3) (Portela and Granell-Canut 2017). Nonetheless, the SDS should admit mistakes without making the user feel responsible for them to maintain user trust (R4) (Branham and Mukkath Roy 2019). For example, the participants are only interested in performing their task and showed no initiative to utilize the small talk function of the open system.
Guided by the DSR paradigm, the primary purpose of this study is to devise and evaluate a design theory for an SDS dialog strategy in customer service. The proposed design theory including 14 requirements and five DPs is informed by the principles of dialog theory (Bunt 2000) and related work in prior conversational agent and SDS research; it is also empirically validated in three iteration rounds through five hypotheses. First, we enrich the body of knowledge by proposing a design theory that provides codifying design knowledge for a class of artifacts (SDS dialog strategies) to address a class of problems according to Walls et al. (1992) and Gregor and Jones (2007). This type of knowledge can be referred to as “nascent design theory,” which provides “knowledge as operational principles/architecture” according to Gregor and Hevner (2013, p. 342). With this contribution, we respond to recent calls for more design knowledge on conversational agents for enhancing user experience in the customer service context in particular (Gnewuch et al. 2017). In the next sections, we discuss the main findings of this study prior to highlighting the major implications for research and practice.
Products and services
On the first call, more assistance should be provided to carefully familiarize users step-by-step with the open system. In particular, the welcome prompt has a significant influence on user expectations; hence, a brief explanation of available self-service options would be useful prior to posing the open question on user intent. By naming the various options, users can easily initiate the desired process and start the conversation without making any mistakes, similar to a closed system. Once users are familiar with the open system (i.e., for subsequent calls or when returning to the main menu), the level of assistance can be reduced.
Read more about https://www.metadialog.com/ here.