Increasingly, firms’ social media (SM) use is on the rise; therefore, effective communication on SM remains a challenge for firms in the digital era. This study explores firms’ different SM engagements with respect to artificial intelligence (AI) and proposes an empirically validated model. A qualitative research design was adopted in which semi-structured interviews were conducted with the chief executive, director, chief entrepreneur, and/or associated top management of firms that use AI. Based on an analysis of these interviews, the study proposes the “microfoundation of social media routine framework” that consists of three processes and four stages. This routine integrates AI manage the engagement of users on firms’ SM. This routine provides the opportunity to establish strong relationships with customers. At the end of the study, we discuss the contributions and implications of the routine and conclude with future research directions.
Implementing decisions with the input of machine learning (ML) helps firms’ optimize their performance. From the perspective of business, numerous studies have been conducted in different domains of ML. Extant literature has emphasized the importance of artificial intelligence (AI) Brynjolfsson and Mcafee, 2017, Ransbotham et al., 2017) and its implications (Belanche et al., 2019, Rodgers et al., 2021) and the importance of big data (Chen et al., 2012, Wiener et al., 2020) and its implications (Chen et al., 2017, Lawrence et al., 2019, Liu, 2019). AI is changing the way people work, live, and solve challenges (Dignum, 2020). Social media (SM) can be managed and developed under the domain of AI. The widespread adoption of SM has created various prospects for gathering appropriate and timely information to improve firms’ operational performance (Castillo, 2016, Imran et al., 2020). Consumers always explore options to reduce the amount of cognitive effort they must apply to decision-making and information-gathering tasks (e.g., Shugan, 1980, Verma et al., 2012). In the past, studies conducted on SM emphasized smooth interaction through social listening and connectedness with stakeholders (e.g., Cao et al., 2021, Istanbulluoglu, 2017). Effective integration of AI and electronic word-of-mouth (E-WOM) on a firm’s SM pages improves communication and management. This study integrated AI and E-WOM to manage communication on firms’ SM pages.
ML is widely used in SM to resolve managerial problems. Van Zoonen and Toni (2016) shared the importance of the coding for SM content through supervised ML. Similarly, Dhaoui et al. (2017) employed ML for SM sentiment analysis to unearth customer issues. Other examples include Cui et al. (2018), who implemented a variety of ML methods to forecast daily sales, and Vermeer et al. (2019), who engaged ML algorithms focused on the relevance of brands and identified the pertinent E-WOM. Some studies have shown positive linkages between ML use and financial performance (Reis et al., 2020). Arora et al. (2020) used the ML techniques of bagging and boosting to train different weak learners (customers) on SM. In the same vein, Liu (2020) adopted ML user-generated content on business-to-business (B2B) firms’ SM and predicted stock performance. Kaiser et al. (2020) employed ML and predicted users’ brand love, brand loyalty, and word-of-mouth endorsement from the content of their brand photos posted on Facebook. Amin et al. (2020) utilized ML to unveil tweets’ topics automatically and reveal financial disclosure tweets. Thus, ML plays an important role in firms’ SM management and the E-WOM of customers.
The rapid growth of information-related technologies has had a huge impact on how, when, and where B2B marketers interact and do business with their customers (Schultz et al., 2012). SM has characteristics of participation, openness, conversation, community, and connectedness (Maresh-Fuehrer and Smith, 2016). Karjaluoto et al. (2015) proposed that people tend to act similarly in different roles on SM. Previous studies have suggested that B2B firms often find it difficult to identify and integrate SM platforms into their digital marketing mix (Iankova et al., 2019, Quinton and Wilson, 2016). Herhausen et al. (2020) conducted a systematic review and survey to find the knowledge gaps in digital marketing. However, the study by Herhausen et al. (2020) further revealed that the majority of firms lack knowledge on SM engagement and use. This gap was further defined as a lack of routines and processes that enable B2B firms to engage in successful SM conversations. Herhausen et al. (2020) named it the “microfoundations of social media use.” To strengthen this area of knowledge, this study proposes to address the gaps and suggests a routine that consists of three processes and four stages through which firms can manage SM engagement.
Jimenez-Marquez et al. (2019) proposed a two-stage framework for SM big data analysis. The first stage is dedicated to data preparation, i.e., aggregating unstructured data and finding an optimal ML model for the data, and the second stage relies on establishing layers of big data architectures focused on getting an outcome from the data by employing most of the ML model from the first stage. This framework manages all the data from SM sources, blogs, chats, and microblogging services. On the other hand, Vermeer et al. (2019) tracked and analyzed E-WOM about brands, products, and services. This study applied “supervised ML” that decides whether seven different types of E-WOM are relevant for a brand and how the firm should be responded to them. Vermeer et al. (2019) proposed the response-worthy E-WOM identification process, but it was limited to the relevance of the E-WOM, type of the E-WOM (unsatisfied, neutral, satisfied), and whether a response to specific E-WOM is required. Their study did not cover the technology-related details in the model. The extant literature covers various methods and aspects of SM use in relation to AI and its benefits in various propositions but lacks a clear framework or standards. To address this research void, this study attempted to answer the following research question:1)
How can AI be integrated with a firm’s SM pages to communicate effectively and satisfy customers by answering queries in the minimum time possible?
Thus, the present study is an AI-integrated effort to provide a model that emphasizes the effective and efficient reply to customer queries on SM on an as-needed basis. To address the above research question, we formulated the following research objectives: a) to explore how AI can be integrated with firms’ SM pages to communicate effectively and ensure customer satisfaction by answering queries in the minimum time possible and b) to corroborate evidence from multiple cases to propose a framework for the “Artificial Intelligence Integrated Routine Model.”
In light of the available literature, our study contributes to the current body of knowledge in the following ways. First, our study proposes a detailed routine and processes for B2B customer engagement on SM (e.g., Facebook, YouTube, Twitter, Instagram). This routine comprises a three-process model consisting of four stages, from customer engagement on SM to the appropriate customer-engagement response. Second, this study covers the knowledge gap of the microfoundations of SM use, as Herhausen et al. coined it in 2020, encourages people and institutions to conduct digital marketing research, and confirms that SM enhances firms’ digital capabilities.
The second section discusses the previous literature on the elements of our proposed routine. This is followed by methodology in the third section and analyses and results in the fourth section. We propose the routine in the fifth section, followed by its contributions and implications. We conclude the paper with recommendations for future research in the final section.