Service robots are projected to have implications at the micro, meso, and macro levels for all key stakeholders in the service environment and within society (Wirtz et al., 2018). While the pandemic has accelerated the relevance of service robots, this exciting future comes with a challenging reality for practitioners when implementing such a disruptive technology within pre-existing service environments. While extant literature has provided some insight as to how service robots might disrupt the relationships within the service triad via their projected roles, the actual challenges and dilemmas created within the service triad have not been investigated (the service trilemma). The term trilemma represents multiple dilemmas between stakeholders of the service triad, as a result of tensions in wellbeing within the traditional symbiotic relationships between the three actor categories. A key factor that can create a service trilemma is the service robot’s social capabilities that make it an entirely new social entity, rather than merely a new technological device. Hence, we need to better understand the social relationships service robots have based on their delivery in the frontline service environment. Thus, this paper seeks to answer the research question, how does the task and contextual performance of service robots challenge the symbiotic relationships within the service triad?

The purpose of this paper is to outline a new framework that provides exploration of the service trilemma, representing the challenges created by service robots to the pre-existing symbiotic relationships in the service triad (customer, frontline employee, and organisation) in the frontline service environment. In this systematic literature review we screened 171 papers, which were pulled from five searches across ProQuest, EBSCOHost, Taylor and Francis, Science Direct, and SpringerLink databases. Exclusion criteria included technical papers, conference articles, non-peer reviewed, non-English, abstract only, non-full papers, prior to 2010, duplicated articles, non-tangible service robots, and journals with minimum (below Q1) impact factors, and was applied following the PRISMA method (Page et al., 2021). We then conducted a systematic review of the 47 remaining articles via a two-stage thematic analysis utilising Nvivo software.

A framework was developed that includes the organisational, frontline employee, and customer motivations and barriers that contribute to each tension that make up the trilemma. The theoretical contributions of this framework is the Robot Service Trilemma, which through a dilemma lens evaluates the challenges service robots bring to the pre-existing relationships within the service environment, as a result of the service robot’s integrated task and contextual performance. These tensions have been found to challenge social wellbeing inclusive of social hierarchical tensions, liability of service failure, and service quality dissonance caused by a service robot. Additionally, this framework provides exciting directions for future research. The managerial implications are relevant to service industries as diverse as medical to restaurant contexts. Thus, this research provides direction for practitioners to mitigate the effects of any resulting tensions the deployment of service robots may cause, and inform long term strategic approach.

Purpose
Chatbots are becoming the first touchpoint in customer service. Recent studies have investigated chatbots' customers' acceptance through chatbots' functional elements using the technology acceptance model (TAM). However, chatbots are limited in terms of relational and emotional capabilities, and there is a dearth of research that studies specifically how chatbots' lack of emotional awareness affects customers' acceptance. Therefore, this paper investigates customer acceptance of chatbots from a relational point of view in agreement with the Service Robots Acceptance Model (sRAM) by Wirtz et al., focusing on two moderators: 1) emotional awareness and 2) technology literacy. Understanding if chatbots’ emotion detection capability enhances the customer-chatbot rapport is crucial for supporting customer engagement during the interaction. The paper also highlights the role of technology literacy in the chatbots' adoption, represented by the customer's previous experience in new technologies.

Design/Methodology/approach
This study examines the motivations to adopt chatbots for customer service based on data collected via a crowdsourcing platform from 301 millennials, analysed using ANOVA methods. The paper employed purposive and convenience sampling techniques to recruit respondents targeting only participants familiar with chatbots.

Findings
Findings show that functional, social, and relational elements drive the adoption of chatbots and reveal the moderating role of emotional awareness and technology literacy. The study contributes to the research on service robots by adding a new relational perspective based on building the customer-chatbot “rapport” through emotional awareness, a concept practically absent from previous empirical studies. Emotional awareness can mediate the effects of negative emotions and moderate the chatbots’ acceptance by strengthening the relational elements of trust and rapport during automated firm-customer encounters. Technology literacy, due to the impact on the functional elements, can be also considered a moderator of the sRAM.

Research limitations/implications
While empirically validating and extending the sRAM for chatbots, this study adds a new perspective regarding the underexplored role of customer-robot rapport building when the chatbot can recognise customer’s emotions during the interaction.
The paper also contributes to research on service frontlines, where emotional awareness in technology-driven encounters has only been implied but not tested. In particular, it highlights that sRAM is a framework that overcomes the limitations of the TAM. The moderator role of emotional awareness on the acceptance of chatbots is certainly the most relevant implication of this study. It also demonstrates that not only the functional elements, such as the cognitive appraisal of chatbots’ perceived usefulness and ease of use, drive the chatbots’ adoption but above all the emotional and relational ones.

Originality/value
The paper contributes to a more holistic understanding of chatbots’ adoption and provides managerial guidance for successfully implementation of such technologies in customer service. This research takes a significant step forward by providing support for a new way of understanding customer acceptance of AI-powered chatbots linked to customer emotions. Service managers cannot fail to consider the impact that emotional awareness has on both direct and indirect customer engagement during automated service encounters.


Keywords: Artificial Intelligence; Chatbots; Emotional Awareness; Technology Literacy, Service Robot Acceptance Model.

Millions of jobs will potentially be lost to robots because they are more reliable, free from human error and fatigue (Huang and Rust, 2018) and may “prevail in the retail, financial, health care, education, transportation, and communication industries.” (Mende, Scott, van Doorn, Grewal, & Shanks, 2019; Ostrom, Fotheringham, & Bitner, 2019, p. 77).

An initial battery of all robots available worldwide was created via internet crawling and studying the world’s largest catalog of robots. A focus group of twenty participants shortlisted services that robots could offer (using government employment lists). This was then transformed into an online questionnaire.

The 480 respondents sample represents a big chunk of the population who will face the brunt of working with co-bots. These are the ages of 19 to 44 at 93% of the sample, 48% having an undergraduate degree and 47% being postgraduates, with most professions being Business, Engineering, Architecture, Education, Computer and Mathematics.

Firstly, respondents had to state, on a scale of 1 to 10 how negative/positive their attitude towards robots was. Two thirds of the respondents had a positive attitude towards robots as service providers (RaSP), whereas 15% were neutral and 18% had a negative attitude.

We proceeded with asking the respondents to state their attitude towards robots providing particular services (27 in total).

The results revealed that in services that are highly mundane (such as house cleaning and delivery) robots will be highly welcomed by humans whereas in services that relate to children or young people (such as babysitting and teaching) and psychologists, robots receive a very high disapproval. This is similar to what Wirtz et. al (2018) name as services that are high in emotional and social tasks.

Professions where high customization is expected (barber/hairdresser, doctor/nurse, lawyer) robots are not very well accepted, neither were the art-professions such as singer etc. This was also the case for police officers and pilots.

Lastly, a video of actual robots employed as service providers was shown. It included three types of robots: a humanoid receptionist, a toy-looking room assistant and an arm luggage storing robot.

Then, the respondents were asked to state which emotions they felt with each robot separately and whether they would use the robot services (as a proxy to purchase intention). The results, were surprising in the sense that the humanoid did not elicit negative emotions such as disgust, anger, fear, sadness or stress, but it did not make the respondents feel warmth or state that that robot would interact effectively with them. On the other hand, the arm luggage sorting robot did not elicit negative emotions but the respondents said that they found it useful, and generally had favorable intentions. The toy-looking room assistant generated the least negative emotions and more positive emotions.

All robot types had a common result: most of the respondents reported that they were neutral to feeling trust towards the specific robot. This valence might change though, as humans interact more and more with robots and any advantages and disadvantages become more evident.

Research has dedicated considerable attention towards determining whether service robots do not only bring organizational benefits (Lu et al., 2020), but whether they also enhance the customer service experience (Choi et al., 2021; Yoganathan et al., 2021). Currently, state-of-the-art research on service robots is in general either of a conceptual nature or uses lab experiments to investigate customer perceptions (Huang and Rust, 2018; Van Doorn et al., 2017; Wirtz et al., 2018). Lu et al (2020) argue that present research on service robots misses out on the social complexity of the real world and call for future field research to understand the real-life impact of frontline service robots. Thus, this paper seeks to answer the research question, how do frontline service robots impact the service experience for hotel guests?

To address this gap, we have analyzed 17,236 reviews by hotel guests for eight different hotels from five leading online customer review websites in the travel and hospitality industries. All eight hotels employ the same service robot. For the collected comments, we analyzed text-based features and related them to the overall service experience, using the self-reported reviewer star rating as a proxy for the service experience. Thus, our approach is very similar to previous text mining studies (e.g., Villarroel Ordenes et al., 2017).

Our preliminary findings show that reviewers that mention the service robot in their comments rate their overall service experience significantly more positively. Intriguingly, when reviewers call the robot by name, this positive relationship is even stronger. This puts forward the question whether service robots positively impact service experiences especially for those reviewers who humanize the robot more, or whether reviewers who rate their experience more positively are more likely to humanize service robots.

As we base our analysis on field data, our study has the potential to offer managers a better understanding of the real-life impact of service robots on the overall service experience. In this way, we facilitate a comprehensive understanding of the empirical reality as reflected in customers’ online generated content. By doing so, we answer to several calls for field research (Odekerken-Schröder et al., 2021; Mende et al., 2019). The contributions of this study are twofold: from a theoretical perspective, we advance the knowledge on personification of service robots, where referring to the robot by name is indicative for personification and positively related to satisfaction with the overall service experience (Purington et al., 2017); our methodological contribution lies in the novelty of text based analysis of online customer review data within the field of service robot research.

Moving forward, we aim to develop a conceptual framework that is grounded in the results of our field study, where we seek to explain the motivational drivers or customer differences underlying the observed relationship (Tausczik and Pennebaker, 2010). To validate the hypothesized relationships, we plan to conduct further review-based analyses to take full advantage of available field data, as well as additional field and lab studies. These studies deepen our understanding of real-life customer-robot interactions in hospitality services.