Conference Paper

Creating Humanlike Chatbots: What Chatbot Developers Could Learn From Webcare Employees In Adopting A Conversational Human Voice

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Currently, conversations with chatbots are perceived as unnatural and impersonal. One way to enhance the feeling of humanlike responses is by implementing an engaging communication style (i.e., Conversational Human Voice (CHV); Kelleher, 2009) which positively affects people’s percep-tions of the organization. This communication style contributes to the effectiveness of online communication between organizations and customers (i.e., webcare), and is of high relevance to chatbot design and development. This project aimed to investigate how insights on the use of CHV in organizations’ messages and the perceptions of CHV can be implemented in customer service automation. A corpus study was conducted to investigate which linguistic elements are used in organizations’ messages. Subsequently, an experiment was conducted to assess to what extent linguistic ele-ments contribute to the perception of CHV. Based on these two studies, we investigated whether the amount of CHV can be identified automatically. These findings could be used to design humanlike chatbots that use a natural and personal communication style like their human conversation partner.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... Besides application area-centered research, researchers try to survey the chatbot ecosystem on a more general business level. Thereby, research directions tackle, e.g., trust aspects, humanizing the chatbot, and challenges [2,[9][10][11][12]. This generalized research can typically be reused and adapted for deviating scenarios or use cases. ...
... To reach these potentials of chatbots at workplaces, a lot of design research has been done so far. Besides the major focus on customer-focused areas, e.g., [7,10,21,22], for example, Elshan and Ebel [23] survey chatbots as teammates, and Winkler et al. [6] apply them for problem-solving in businesses. Besides this, chatbots were also used as a means for feedback exchange [24]. ...
... In addition, our workshop participants pointed out, that in this step also the desired level of humanity and anthropomorphism must be clarified 21 . Hereto, enterprises can already rely on a large research stream, e.g., [9,10,28,29,40]. This also encompasses the definition of chatbots' persona 22 [16], e.g., conversation style, appearance, or name. ...
... Additionally, the dialog design should encompass multimedia elements to enrich the dialog (Bittner and Shoury 2019). Furthermore, general aspects of humanized chatbots should also be taken into account to foster user acceptance, e.g., typing delays, gender, or phrasing (Gnewuch et al. 2018;Liebrecht and van Hooijdonk 2020). The chatbot should be a web-based application to provide the functionalities device-and location-independently at any time, and to address the fundamental chatbot characteristics (Meyer von Wolff et al. 2020). ...
... Also, consideration should be given to including support for small talk, since this can further enhance the adoption of chatbots in educational settings, or further in the support department (Hobert and Berens 2020). This was also already found by (Liebrecht and van Hooijdonk 2020), who try to enhance the natural and personal feeling of a chatbot. Nevertheless, our results and evaluations already show a basic acceptance of chatbots in IT-support tasks. ...
Article
Understanding the user experience of chatbots for customer service is essential to realize the potential of this technology. Such chatbots are typically designed for efficient and effective interactions, accentuating pragmatic quality, and there is a need to understand how to make these more pleasant and engaging, strengthening hedonic quality. One promising approach is to design for more humanlike chatbot interactions, that is, interactions resembling those of skilled customer service personnel. In a randomized experiment (n = 35) we investigated two chatbot interaction design features that may strengthen the impression of a humanlike character: (a) topic-led conversations, encouraging customer reflection, in contrast to task-led conversations, aiming for efficient goal completion, and (b) free text interaction, where users interact mainly using their own words, rather than button interaction, where users mainly interact through predefined answer alternatives. Dependent variables were participant perceptions of anthropomorphism and social presence, two key concepts related to chatbot human likeness, in addition to pragmatic quality and hedonic quality. To further explore user perceptions of the interaction designs, the study also included semi-structured interviews. Topic-led conversations were found to strengthen anthropomorphism and hedonic quality. A similar effect was not found for free text interaction, reportedly due to lack in chatbot flexibility and adaptivity. Implications for theory and practice are suggested.
Article
The conversational human voice (CHV) is an extensively studied and adopted communication style in online brand communication. However, in previous research the way in which CHV is operationalized differs considerably: the type and the number of linguistic elements used to establish a sense of CHV in online brand messages varies. Moreover, it is still unknown how CHV operationalizations contribute to consumers’ perceptions of CHV, which consequently could affect their evaluation regarding the message and the brand. In this paper, we addressed these issues by conducting an integrative literature review and a perception experiment, and consequently present a taxonomy of linguistic elements related to message personalization, informal speech, and invitational rhetoric that can be used to operationalize CHV systematically in future studies in online brand communication. Directions for future research and managerial implications are discussed.
Article
Full-text available
Conversational agents (CAs) are software-based systems designed to interact with humans using natural language and have attracted considerable research interest in recent years. Following the Computers Are Social Actors paradigm, many studies have shown that humans react socially to CAs when they display social cues such as small talk, gender, age, gestures, or facial expressions. However, research on social cues for CAs is scattered across different fields, often using their specific terminology, which makes it challenging to identify, classify, and accumulate existing knowledge. To address this problem, we conducted a systematic literature review to identify an initial set of social cues of CAs from existing research. Building on classifications from interpersonal communication theory, we developed a taxonomy that classifies the identified social cues into four major categories (i.e., verbal, visual, auditory, invisible) and ten subcategories. Subsequently, we evaluated the mapping between the identified social cues and the categories using a card sorting approach in order to verify that the taxonomy is natural, simple, and parsimonious. Finally, we demonstrate the usefulness of the taxonomy by classifying a broader and more generic set of social cues of CAs from existing research and practice. Our main contribution is a comprehensive taxonomy of social cues for CAs. For researchers, the taxonomy helps to systematically classify research about social cues into one of the taxonomy's categories and corresponding subcategories. Therefore, it builds a bridge between different research fields and provides a starting point for interdisciplinary research and knowledge accumulation. For practitioners, the taxonomy provides a systematic overview of relevant categories of social cues in order to identify, implement, and test their effects in the design of a CA.
Conference Paper
Full-text available
Text messaging-based conversational agents (CAs), popularly called chatbots, received significant attention in the last two years. However, chatbots are still in their nascent stage: They have a low penetration rate as 84% of the Internet users have not used a chatbot yet. Hence, understanding the usage patterns of first-time users can potentially inform and guide the design of future chatbots. In this paper, we report the findings of a study with 16 first-time chatbot users interacting with eight chatbots over multiple sessions on the Facebook Messenger platform. Analysis of chat logs and user interviews revealed that users preferred chatbots that provided either a 'human-like' natural language conversation ability, or an engaging experience that exploited the benefits of the familiar turn-based messaging interface. We conclude with implications to evolve the design of chatbots, such as: clarify chatbot capabilities, sustain conversation context, handle dialog failures, and end conversations gracefully.
Conference Paper
Full-text available
The idea of interacting with computers through natural language dates back to the 1960s, but recent technological advances have led to a renewed interest in conversational agents such as chatbots or digital assistants. In the customer service context, conversational agents promise to create a fast, convenient, and cost-effective channel for communicating with customers. Although numerous agents have been implemented in the past, most of them could not meet the expectations and disappeared. In this paper, we present our design science research project on how to design cooperative and social conversational agents to increase service quality in customer service. We discuss several issues that hinder the success of current conversational agents in customer service. Drawing on the cooperative principle of conversation and social response theory, we propose preliminary meta-requirements and design principles for cooperative and social conversational agents. Next, we will develop a prototype based on these design principles.
Chapter
Full-text available
When consumers are dissatisfied with a consumption experience, they usually respond in one of the following ways (Hirschman, 1970): (1) stop using an organization’s products/services and take their business to a competitor, (2) file a complaint with the organization that is responsible for the dissatisfying consumption experience, or (3) talk about their dissatisfying consumption experience with fellow consumers (negative word of mouth). Janelle McCoy, a former loyal customer of Chevrolet, decided to combine all responses with the help of social media. In a series of comments on Facebook and Twitter, in which Chevrolet was either tagged or addressed (@chevrolet), Janelle shared her dissatisfaction with one of Chevrolet’s car dealers. In doing so, she not only engaged in negative word of mouth (NWOM) but also complaint behaviour. As can be seen from the excerpt of the Twitter dialogue depicted by Figure 4.1, Janelle’s comments addressed a double audience consisting of not only other consumers but also the organization responsible for the dissatisfying consumption experience. Consumers such as Janelle increasingly voice their complaints as electronic NWOM, with the aim to draw the attention of organizations and, as such, enforce service excellence. Thus, after receiving no satisfactory response from Chevrolet, Janelle decided to take her business to a competitor and to share this decision with other consumers on Facebook and Twitter as well.
Conference Paper
Full-text available
In this paper we compare the effects of different organizational crisis responses in social media after negative publicity regarding the organization. Building on earlier work in crisis communication (e.g., Bradford & Garrett, 1995; Kim, Avery & Lariscy, 2009), we distinguish apology and denial as organizational responses. Building on recent work on corporate blogs by Kelleher (2009), we distinguish a personal and a corporate tone of voice and predict that a personal tone of voice is more effective in social media because it makes the organization look more human and more committed to the relationship with its customers. The hypotheses were tested using a 2 (apologies vs. denial) x 2 (corporate vs. personal) + 1 control group (no response) between subjects experimental design, building on a recent case in which H&M got involved in a discussion with its Facebook community after reports about cutting up and throwing away unsold garments (instead of donating the garments to the homeless). The 125 participants read an online news report about the incident, and were then shown the H&M Facebook page on which H&M either apologized or denied the incident, and did so in a personal or a corporate manner. A control group only read the online news story. The results showed that apologizing led to higher credibility and a more positive attitude towards the response, but increased the perceived responsibility of the organization for the crisis event. A personal tone of voice increased conversational human voice, communicated relational commitment and decreased negative cognitive responses, especially in the case of denial. The results are discussed in the light of the difficulties that organizations face when operating in social media.
Article
Full-text available
Most traditional rhetorical theories reflect a patriarchal bias in the positive value they accord to changing and thus dominating others. In this essay, an alternative rhetoric—invitational rhetoric—is proposed, one grounded in the feminist principles of equality, immanent value, and self‐determination. Its purpose is to offer an invitation to understanding, and its communicative modes are the offering of perspectives and the creation of the external conditions of safety, value, and freedom.
Article
Full-text available
Web 2.0 has empowered consumers to voice complaints with reduced costs (physical and psychological), and to share these with a multitude of other consumers on the Internet. As a public phenomenon, online complaints have a negative impact on consumers' evaluations of brands that are under attack in online complaints. By means of an experiment, we study the most effective means for companies to counter complaints as expressed in negative electronic word of mouth (NWOM). The results show that negative brand evaluations engendered by NWOM can be attenuated by webcare interventions dependent on type of strategy (proactive vs. reactive) and platform used (consumer-generated vs. brandgenerated blog). This effect appeared to be mediated by conversational human voice. The findings are discussed in the light of practical implications for online complaint management.
Article
Full-text available
Abstract This study examined how creating a human presence in organizational online communication affects organization-public relationships and publics' favorable behavioral intentions to engage in word-of-mouth (WOM) and dialogic communications. Four hypotheses were tested in the context of Twitter through a 2×2 (presence: human vs. organizational×organization type: nonprofit vs. for-profit) within-subjects design. The results revealed that conversational human voice was perceived to be higher for Twitter pages of organizations with a human presence than for those with an organizational presence. Providing a human presence on social media through the use of social media managers' avatars and names appeared to promote favorable organization-public relationships and positive WOM communication. However, dialogic communication intentions did not significantly differ between organizations incorporating a human presence versus an organizational presence into their Twitter pages. The proposed dynamic role of human presence versus organizational presence adds a new perspective as to how organizations can take better advantage of interpersonal aspects of social media.
Article
Full-text available
A chatbot is a machine conversation system which interacts with human users via natural conversational language. Software to machine-learn conversational patterns from a transcribed dialogue corpus has been used to generate a range of chatbots speaking various languages and sublanguages including varieties of English, as well as French, Arabic and Afrikaans. This paper presents a program to learn from spoken transcripts of the Dialogue Diversity Corpus of English, the Minnesota French Corpus, the Corpus of Spoken Afrikaans, the Qur'an Arabic-English parallel corpus, and the British National Corpus of English; we discuss the problems which arose during learning and testing. Two main goals were achieved from the automation process. One was the ability to generate different versions of the chatbot in different languages, bringing chatbot technology to languages with few if any NLP resources: the corpus-based learning techniques transferred straightforwardly to develop chatbots for Afrikaans and Qur'anic Arabic. The second achievement was the ability to learn a very large number of categories within a short time, saving effort and errors in doing such work manually: we generated more than one million AIML categories or conversation-rules from the BNC corpus, 20 times the size of existing AIML rule-sets, and probably the biggest AI Knowledge-Base ever.
Article
Full-text available
This paper investigates the linguistic worth of current ‘chatbot’ programs – software programs which attempt to hold a conversation, or interact, in English – as a precursor to their potential as an ESL (English as a second language) learning resource. After some initial background to the development of chatbots, and a discussion of the Loebner Prize Contest for the most ‘human’ chatbot (the ‘Turing Test’), the paper describes an in-depth study evaluating the linguistic accuracy of a number of chatbots available online. Since the ultimate purpose of the current study concerns chatbots' potential with ESL learners, the analysis of language embraces not only an examination of features of language from a native-speaker's perspective (the focus of the Turing Test), but also aspects of language from a second-language-user's perspective. Analyses indicate that while the winner of the 2005 Loebner Prize is the most able chatbot linguistically, it may not necessarily be the chatbot most suited to ESL learners. The paper concludes that while substantial progress has been made in terms of chatbots' language-handling, a robust ESL ‘conversation practice machine’ (Atwell, 1999) is still some way off being a reality.
Article
Full-text available
Organizations face unique challenges in communicating interactively online with publics that comprise dauntingly large numbers of individuals. This online survey examined the perceptions of people who had experienced interactive communication with a large consumer-tech-industry company via organizational blogs. Those reporting the greatest exposure to the blogs in this study were more likely to perceive the organization as communicating with a conversational voice. Conversational human voice and communicated relational commitment (relational maintenance strategies) correlated positively with trust, satisfaction, commitment, and control mutuality (relational outcomes). Building on prior research, this survey supports a model of distributed public relations-one in which key outcomes of public relations are fostered by a wide range of people communicating interactively while representing an organization.
Conference Paper
Full-text available
This paper presents a new experimental paradigm for the study of human-computer interaction, Five experiments provide evidence that individuals' interactions with computers are fundamentally social. The studies show that social responses to computers are not the result of conscious beliefs that computers are human or human-like. Moreover, such behaviors do not result from users' ignorance or from psychological or social dysfunctions, nor from a belief that subjects are interacting with programmers. Rather, social responses to computers are commonplace and easy to generate. The results reported here present numerous and unprecedented hypotheses, unexpected implications for design, new approaches to usability testing, and direct methods for verii3cation.
Article
Full-text available
This study develops and tests operational definitions of relational maintenance strategies appropriate to online public relations. An experiment was designed to test the new measures and to test hypotheses evaluating potential advantages of organizational blogs over traditional Web sites. Participants assigned to the blog condition perceived an organization’s “conversational human voice” to be greater than participants who were assigned to read traditional Web pages. Moreover, perceived relational strategies (conversational human voice, communicated relational commitment) were found to correlate significantly with relational outcomes (trust, satisfaction, control mutuality, commitment).
Article
Mensachtige chatbots: een zegen voor online sociale klantendienst? Chatbots voor klantenservice gaan snel vooruit. Er was echter weinig bekend over de communicatiestijl die merken zouden moeten gebruiken in chatbotcommunicatie. Uit de literatuur over organisatiecommunicatie tussen merken en klanten werd het concept van een menselijke stem (Kelleher, 2009; Kelleher & Miller, 2006) opgenomen in chatbotcommunicatie. In een 2x2 experiment tussen proefpersonen gingen 125 deelnemers in gesprek met een chatbot voor klantenservice in de reissector. Naast de aanwezigheid of afwezigheid van een menselijke stem, verschilden de chatbots ook in merkbekendheid. De resultaten onthulden klantenservicechatbots met een gemoedelijke menselijke stem die de merkattitude en de waargenomen warmte verhoogt via sociale aanwezigheid, zonder de waargenomen merkcompetentie aan te tasten. Bovendien hebben we geen bewijs gevonden voor merkbekendheid om deze bevindingen te matigen. Onze resultaten ondersteunen de theorie van Computers zijn sociale actoren (CASA) en geven aan dat de waargenomen menselijkheid van computers zou kunnen toenemen door de aangenomen communicatiestijl.
Article
“Annoying that the bike hasn’t been removed! Can you give me a case number? ^EK”: Conversational Human Voice in webcare of Dutch municipalities The conversational communication style organisations use in webcare is an important factor affecting its success, and is referred to as the Conversational Human Voice (CHV, Kelleher, 2009; Kelleher & Miller, 2006). This communication style reflects attributes such as treating others as humans, using a personal communication style, and being open to dialogue. Although several experimental studies investigated the relation between conversational linguistic elements in webcare messages and perceived CHV (for example Crijns, Cauberghe, Hudders & Claeys, 2017; Huibers & Verhoeven, 2014), there are considerable differences in the type and number of linguistic elements they used. Therefore we developed an instrument based on scientific research to identify conversational linguistic elements reliably. Next, we investigated how often these elements occurred in a corpus of 480 webcare conversations between twenty Dutch municipalities and their citizens, and to what extent they differ in its usage. The results showed that the identification instrument was reliable. Furthermore, municipalities often personalize their webcare responses, but hardly use informal language and invitational rhetoric. Large municipalities, such as Amsterdam and ’s-Hertogenbosch, appeared to be more progressive in applying conversational elements in webcare than medium-sized municipalities, such as Gouda and Deventer.
Article
Disembodied conversational agents in the form of chatbots are increasingly becoming a reality on social media and messaging applications, and are a particularly pressing topic for service encounters with companies. Adopting an experimental design with actual chatbots powered with current technology, this study explores the extent to which human-like cues such as language style and name, and the framing used to introduce the chatbot to the consumer can influence perceptions about social presence as well as mindful and mindless anthropomorphism. Moreover, this study investigates the relevance of anthropomorphism and social presence to important company-related outcomes, such as attitudes, satisfaction and the emotional connection that consumers feel with the company after interacting with the chatbot.
Article
Social media platforms enable firms to communicate directly and often publicly with individual consumers. In this research, comprising four online studies, the authors investigate how the tone of voice used by firms (human vs. corporate) influences purchase intentions on social media. Findings suggest that a human tone of voice is not always the firm's best option. Study 1a (N = 174) shows that using a human voice, instead of the more traditional corporate voice, can increase a consumer's hedonic value on social media and also purchase intentions. However, that influence of a human voice on purchase intentions is stronger when the consumer is looking at a brand page with a hedonic goal in mind (versus a utilitarian one). Study 1b (N = 342) shows that the presence of several negative comments about a brand on social media acts as a boundary condition, nullifying the influence of a human voice on purchase intentions. Studies 2a (N = 154) and 2b (N = 202) show in different settings that using a human voice can even reduce purchase intentions in contexts of high situational involvement, due to perceptions of risk associated with humanness. The results contribute to the literature surrounding the effects of conversational human voice, while also providing managers with a set of guidelines to help inform and identify which tone of voice is best adapted to each communications scenario.
Article
Purpose Social media encourage interactions between customers and brands. Concerning the cues utilized during social media interactions, verbal cues (i.e. the language used) gain importance, since non-verbal and paraverbal cues are hard to convey via social media. Looking at interpersonal interactions, interlocutors adopt each other’s language styles or maintain their own language style during interactions to build trust. Transferring these insights to social media, the purpose of this paper is to test the effects of a brand’s language style accommodation in brand-customer interactions on brand trust and on its antecedents. Design/methodology/approach Two quantitative pre-studies (n1 (questionnaire)=32, n2 (laboratory experiment)=199), and one quantitative main study (n3 (laboratory experiment)=427) were conducted to determine the effects of a brand’s language style accommodation on brand trust. Findings In line with communication accommodation theory, this paper reveals that the impact of a brand’s accommodation strategy on brand trust is mediated by perceived relationship investments, such as perceived interaction effort, benevolence, and quality of interaction. This paper also underscores language style’s roles and its fit, and sheds light on situational factors such as purchase decision involvement and the valence of the content. Originality/value This paper is the first to transfer cross-disciplinary theories on interpersonal interactions to brand-customer interactions in social media. Thus, the authors derive the effects of language style accommodation on brand trust as well as further mediating effects.
Conference Paper
The past four years have seen the rise of conversational agents (CAs) in everyday life. Apple, Microsoft, Amazon, Google and Facebook have all embedded proprietary CAs within their software and, increasingly, conversation is becoming a key mode of human-computer interaction. Whilst we have long been familiar with the notion of computers that speak, the investigative concern within HCI has been upon multimodality rather than dialogue alone, and there is no sense of how such interfaces are used in everyday life. This paper reports the findings of interviews with 14 users of CAs in an effort to understand the current interactional factors affecting everyday use. We find user expectations dramatically out of step with the operation of the systems, particularly in terms of known machine intelligence, system capability and goals. Using Norman's 'gulfs of execution and evaluation' [30] we consider the implications of these findings for the design of future systems.
Article
By performing tasks traditionally fulfilled by service personnel and having a humanlike appearance, virtual customer service agents bring classical service elements to the web, which may positively influence customer satisfaction through eliciting social responses and feelings of personalization. This paper sheds light on these dynamics by proposing and testing a model drawing upon the theories of implicit personality, social response, emotional contagion, and social interaction. The model proposes friendliness, expertise, and smile as determinants of social presence, personalization, and online service encounter satisfaction. An empirical study confirms the cross-channel applicability of friendliness and expertise as determinants of social presence and personalization. Overall, the study underlines that integration between technology and personal aspects may lead to more social online service encounters.
Article
Following Langer (1992), this article reviews a series of experimental studies that demonstrate that individuals mindlessly apply social rules and expectations to computers. The first set of studies illustrates how individuals overuse human social categories, applying gender stereotypes to computers and ethnically identifying with computer agents. The second set demonstrates thatpeople exhibit overlearned social behaviors such as politeness and reciprocity toward computers. In the third set of studies, premature cognitive commitments are demonstrated: A specialist television set is perceived as providing better content than a generalist television set. A final series of studies demonstrates the depth of social responses with respect to computer ‘personality.’ Alternative explanations for these findings, such asanthropomorphism and intentional social responses, cannot explain the results. We conclude with an agenda for future research.
Article
Empirical studies have repeatedly shown that autonomous artificial entities, so-called embodied conversational agents, elicit social behavior on the part of the human interlocutor. Various theoretical approaches have tried to explain this phenomenon: According to the Threshold Model of Social Influence (Blascovich et al., 2002), the social influence of real persons who are represented by avatars will always be high, whereas the influence of an artificial entity depends on the realism of its behavior. Conversely, the Ethopoeia concept (Nass & Moon, 2000) predicts that automatic social reactions are triggered by situations as soon as they include social cues. The presented study evaluates whether participants´ belief in interacting with either an avatar (a virtual representation of a human) or an agent (autonomous virtual person) lead to different social effects. We used a 2 × 2 design with two levels of agency (agent or avatar) and two levels of behavioral realism (showing feedback behavior versus showing no behavior). We found that the belief of interacting with either an avatar or an agent barely resulted in differences with regard to the evaluation of the virtual character or behavioral reactions, whereas higher behavioral realism affected both. It is discussed to what extent the results thus support the Ethopoeia concept.
An investigation of conversational agent relevance, presence, and engagement
  • R M Schuetzler
  • G M Grimes
  • J S Giboney
Schuetzler, R.M., Grimes, G.M., Giboney, J.S.: An investigation of conversational agent relevance, presence, and engagement. In: Proceedings of Americas' Conference on Information Systems (2018)
Iterative development and evaluation of a social conversational agent
  • A Silvervarg
  • A Jönsson
Silvervarg, A., Jönsson, A.: Iterative development and evaluation of a social conversational agent. In: 6th International Joint Conference on Natural Language Processing, pp. 1223-1229 (2013)