Article Preview
Top1. Introduction
Marketing is a complex decision-making discipline that involves not only the commonly known 4Ps (product, price, promotion, and place) but also strategic issues such as new product development (NPD), customer relationship management (CRM), selling strategies, market segmentation, positioning and targeting, international marketing, marketing research, etc. (Rutz & Watson, 2019). With the ever-increasing amount and importance of “big data,” now scholars are interested in whether appropriate decision-making technologies can solve marketing problems.
Artificial intelligence (AI), which refers to machines and software that exhibit human intelligence, can provide great opportunities to facilitate decision-making in marketing. The existence of AI could be traced back to 1955 when John McCarthy coined the term Artificial Intelligence. In his work, AI was defined as “making a machine behave in ways that would be called intelligent if a human were so behaving” (McCarthy et al., 1955). Since then, AI definition has evolved as “manifested by machines that exhibit aspects of human intelligence” (Huang & Rust 2018, p. 155). With the tools of machine learning, deep learning, neural networks, and natural language processing, AI could “interpret external data correctly, learn from such data, and exhibit flexible adaptation” (Kaplan & Haenlein, 2019, p. 17). There are substantial AI-based examples (Kumar et al., 2016, Huang & Rust, 2017). For instance, when you post a photo on Facebook, it can automatically recognize you and your friends' faces and tag their names; when you are off, Amazon's Alexa and Google-Home work as virtual assistants to take care of your home, set the room temperature, manage your schedules, and control the lights; when you ask Amazon for return, its chatbot serves you all the time.
In this regard, the definition of AI in marketing could be illustrated by those business recognitions. AI can implement simple marketing transactions, such as translating emails or phone calls to automate replies, reading customers’ online comments, and smart retailing by recommending products to customers. AI can leverage machine learning tools to analyze large volumes of customer digital footprints, including reviews, video, images, subscriptions, browsing history, webpage activities, and even facial expression data. Those analyses empowered AI to gain a deep understanding of customers’ preferences, behaviors, likes/dislikes, trends, etc. AI could engage customers through real-time interaction and customized digital advertising recommendations. This fast movement allows AI to stay actively engaged with customers to influence their decision-making.
Due to the ever-increasing interest and importance of AI applications in our marketing field, a comprehensive review that precisely analyzes the AI-marketing (AIM) interface literature is imperative to properly understand the constantly growing AIM field (Siau & Yang, 2017). The literature review aims to provide a holistic view and meaningful research questions in AIM studies. Bibliometric study, social network analysis, main path analysis, and content analysis are effective tools used to conduct this systematic literature review.
Bibliometric studies explore a large amount of content in academic journals, including the journal citation information, to identify its leading trend (Bonilla et al., 2015). Journal authors, keywords, affiliations, and citations are also collected and traced in a bibliometric study. Social network analysis is usually followed after a bibliometric study to provide insights about institutions, countries, and keyword collaboration networks. Social network analysis monitors and interprets social ties among social nodes to visualize how the relationships in a group relate to each other and determine the types of relationships that lead to effective outcomes (Stangor, 2015). The main path analysis is based on social network analysis to detect the meaningful and traceable main paths representing the journals in the social network. The importance of path is measured by “counting the number of times a citation link has been traversed if one exhausts the search from a set of starting nodes to another set of ending nodes” (Hummon & Coreian 1989, p. 50).