Technological advancements are significant drivers of economic development, fostering efficiency gains and creating new markets. Voice assistants (VAs), speech-driven interaction systems integrating Artificial Intelligence (AI), are rapidly gaining adoption in private households through smart speakers. Their potential for generating substantial revenue from innovative user services is significant, with high adoption rates projected due to efficiency gains in managing household systems and devices. However, the sheer volume and disorganized nature of existing research across Computer Science (CS), Social Science (SS), and Business and Management Science (BMS) domains presents challenges for predicting future technology use cases and developing effective strategies. While CS research focuses on technological feasibility, SS explores social acceptance and risks (e.g., bias, privacy), and BMS examines market viability and business models. This paper aims to synthesize this dispersed knowledge from an interdisciplinary perspective, focusing on studies published before May 2020, to identify opportunities and guide future research aligned with user needs and societal desires, encompassing feasibility, viability, and desirability dimensions of innovation.

The paper acknowledges the absence of systematic reviews focusing solely on VAs from a single discipline's perspective. It notes some existing explorations of VAs in healthcare and industry, highlighting critical points such as limitations in conversation continuity and the need for robust voice recognition algorithms. The authors highlight the need to organize the scattered research to inform strategies for developing VA solutions matching private household needs, emphasizing the integration of technical, social, and managerial aspects.

The study employs a systematic literature review combining bibliometric and qualitative content analysis. The bibliometric analysis, using the Scopus database and VOSviewer software, examines co-occurrence networks of keywords to identify thematic clusters within each discipline (CS, SS, BMS). The initial search yielded a large number of articles, which were then screened for relevance. The keywords "voice assistant" and its synonyms, along with "home" and its synonyms were used in the search. The study followed the PRISMA guidelines for systematic reviews. The qualitative content analysis involved three researchers independently assigning articles to thematic clusters, followed by collaborative refinement and discussion to ensure consistency. The process of data cleaning included removing duplicates and irrelevant articles. The final dataset consisted of 207 articles, covering CS (147 articles), SS (47 articles), and BMS (13 articles). Co-occurrence networks were generated to visualize common knowledge patterns and thematic links. The analysis identifies nine thematic clusters, which were further consolidated into four interdisciplinary research streams: Conceptual foundation of VA research, Systemic challenges, enabling technologies and implementation, Efficiency, and VA applications and (potential) use cases.

The analysis reveals nine thematic clusters: Smart devices, Human-computer interaction (HCI) and user experience (UX), Privacy and technology adoption, VA marketing strategies, Technical challenges in VA applications development, Potential future VAs and augmented reality (AR) applications and developments, Efficiency increase by VA use, VAs providing legal evidence, and VAs supporting assisted living. These clusters are then consolidated into four interdisciplinary research streams. The study finds that CS research predominantly focuses on technological advancements and application development, while SS explores user perceptions, privacy concerns, and social impacts. BMS research primarily investigates marketing strategies, business models, and efficiency gains. However, cross-disciplinary collaboration is limited. The analysis reveals that while technological development is advancing, research lacks sufficient integration across the three disciplines. The US leads in research output in both CS and SS, indicating a concentrated focus in this region. The study identifies several limitations and shortcomings of VA solutions, including concerns around data security and passive listening, which are significant barriers to broader adoption of more sophisticated VA applications. Existing VA usage is primarily for simple tasks. The study emphasizes that companies should not overestimate the potential for quick returns on investment in VA technology and that only further integration of AI-enabled services is expected to be a game-changer. User acceptance and trust are key factors driving technology adoption, indicating a need for user-centric product development. The study emphasizes the importance of considering the three dimensions of feasibility, viability, and desirability in developing VA-related business models. The paper highlights the significance of interdisciplinary collaboration to integrate research findings and develop strategies around VA solutions that effectively meet the needs of private households.

The findings highlight the need for integrating domain-specific efforts from CS, SS, and BMS to advance user adoption of complex VA applications. The study proposes three sub-propositions related to interdisciplinary research in the areas of medical care solutions, smart home system efficiency and regulation, and social/economic conditions for VA adoption. The study also advocates for interdisciplinary research to address ecosystem-related challenges to technology adoption, emphasizing the need to investigate changes in regulations, insurance, and real estate aspects impacting VA integration. A conceptual framework is presented, outlining avenues for future interdisciplinary research to address the identified challenges. The study identifies potential business opportunities in smart home systems, assisted living and medical home therapy, and digital forensics, highlighting the need for companies to develop a deep understanding of the potential design of future ecosystems and to create business models that incorporate user needs, privacy concerns, and data security.

This research reveals a critical need for increased interdisciplinary collaboration in the study of VAs in private households. The fragmented nature of current research hinders the development of user-centric and commercially successful applications. The proposed framework guides future research towards integrating findings from CS, SS, and BMS, emphasizing the importance of addressing user privacy concerns, developing robust security measures, and creating user-friendly applications. Future research should explore the impact of factors such as the Covid-19 pandemic and consider employing advanced bibliometric analysis techniques for deeper insights.

The study's limitations include the subjective nature of qualitative analysis and the restriction to articles from the Scopus database, potentially excluding relevant research from other sources. The focus on three scientific domains and data collected before May 2020 limits the inclusion of recent research and the impact of events such as the COVID-19 pandemic. Future research could address these limitations by expanding the databases considered and incorporating more recent studies to comprehensively capture the evolving landscape of VA research.