A Framework for Collaborative Artificial Intelligence in Marketing

Highlights

• AI advances from mechanical to thinking to feeling, changing how AI should be used.

• AI and human intelligence (HI) complement best as collaborative teams.

• Lower-level AI augments higher-level HI.

• AI first augments and then replaces HI at a given intelligence level.

• Move HI to a higher intelligence level when AI automates the lower level.

Abstract

We develop a conceptual framework for collaborative artificial intelligence (AI) in marketing, providing systematic guidance for how human marketers and consumers can team up with AI, which has profound implications for retailing, which is the interface between marketers and consumers. Drawing from the multiple intelligences view that AI advances from mechanical, to thinking, to feeling intelligence (based on how difficult for AI to mimic human intelligences), the framework posits that collaboration between AI and HI (human marketers and consumers) can be achieved by 1) recognizing the respective strengths of AI and HI, 2) having lower-level AI augmenting higher-level HI, and 3) moving HI to a higher intelligence level when AI automates the lower level. Implications for marketers, consumers, and researchers are derived. Marketers should optimize the mix and timing of AI-HI marketing team, consumers should understand the complementarity between AI and HI strengths for informed consumption decisions, and researchers can investigate innovative approaches to and boundary conditions of collaborative intelligence.

Introduction

In recent years, AI has moved retailing forward in many ways, such as making big data available for prediction, facilitating more informed retail and consumption decisions, enabling visual display and merchandise, and creating customer engagement (Grewal, Roggeveen, and Nordfalt 2017). AI also is rapidly gaining popularity and importance in the more general marketing area. On the practice side, many marketing functions¹ have deployed AI applications, such as robots for consumer greeting, big data analytics for price adjustment and prediction, recommender systems for product and promotional personalization, natural language processing for customer engagement and in-store experience optimization, and sentiment analysis for customer satisfaction tracking, among others.

On the academic side, we similarly observe the proliferation of AI research in marketing. For example, the impact of in-store AI on retailing (Grewal, Roggeveen, and Nordfalt 2017; Grewal et al. 2020), the investigation of the psychological and cultural barriers to consumer adoption of autonomous shopping systems (de Bellis and Venkataramani Johar 2020), the development of explainable automated product recommendation methods (Marchand and Marx 2020), the application of deep convolutional neural networks to forecast retail sales (Ma and Fildes 2021), the use of big data and unstructured data in marketing (e.g. Balducci and Marinova 2018; Grewal, Roggeveen, and Nordfalt 2017; Wedel and Kannan 2016); the applications of various machine learning methods, such as video mining (Li, Shi, and Wang 2019), text analysis (Berger et al., 2019, Humphreys and Wang, 2018), topic modeling (Antons and Breidbach 2018), semantic analysis (Liu and Toubia 2018), dynamic online pricing (Misra, Schwartz, and Abernethy 2019); and the impact of AI applications, such as adaptive personalization for music and news (Chung, Rust, and Wedel 2009; Chung, Wedel, and Rust 2016), IoT and consumption experience (Hoffman and Novak, 2018, Novak and Hoffman, 2019), AI’s impact on consumer experiences (Puntoni et al. 2020), on service (Huang and Rust 2018), and on the economy (Huang, Rust, and Maksimovic 2019; Rust and Huang 2021), and the strategic use of AI to engage customers (Huang and Rust 2021a).

Although the importance of AI has been well-recognized in marketing practice and research, there are persistent debates about whether AI augments or replaces humans (Huang and Rust, 2018, Davenport and Kirby, 2015, Frey and Osborne, 2017, Malone, 2018). Reflected in marketing studies, some studies find support for replacement, such as consumer perceived warmth and liking of anthropomorphized robots (Kim, Schmitt, and Thalmann 2019) and anthropomorphized display of product and consumer’s attachment and willingness to pay (Yuan and Dennis 2019), while some studies find support for augmentation, such as the contrasts between frontline employees and AI with respect to dimensions of service training and learning, customer experience, firm strategy, and impact on society (Wirtz et al. 2018) and consumer resistance to using medical AI, with the result of needing human mediation (Longoni, Bonezzi, and Morewedge 2019).

These disagreements generate persistent concerns about the role and impact of AI in marketing, especially regarding to what extent AI should be used to perform marketing tasks in replacement of human marketers, and in what way AI should be used by marketers and consumers for augmentation, to avoid unnecessary replacement. Previous studies have established that there are multiple AI intelligences (Huang and Rust 2018; Huang, Rust, and Maksimovic 2019; Rust and Huang 2021) that can provide multiple benefits in service to engage customers (Huang and Rust 2021a). Huang and Rust (2018) first proposed the novel multiple AI intelligences view, supported by analytical theory, that AI advances from mechanical, to thinking, to feeling intelligence (based on how difficult for AI to mimic human intelligences). Huang, Rust, and Maksimovic (2019) provide empirical evidence based on the U.S. government data for the multiple AI intelligences view and their impacts on the economy, predicting that when AI does the thinking tasks and jobs, humans need to upgrade to feeling tasks and jobs. Huang and Rust (2021a) illustrate how multiple AI intelligences, individually and collectively, can be used to engage customers for different service benefits at different service stages, including standardization, personalization, and relationalization. Huang and Rust (2021b) provide a strategic framework using the multiple benefits of AI for marketing research, marketing segmentation, targeting, and positioning (STP), and marketing actions (4P/4C).

Building on these existing studies, we develop a general framework for collaborative intelligence in marketing to address the important issue of how to use AI to augment marketers and consumers at different intelligence levels. The framework asserts that 1) AI has relative strengths over HI (human intelligences of marketers and consumers) for performing mechanical and analytical marketing and consumption tasks, whereas HI (currently) has relative strengths over AI for performing contextual, intuitive, and feeling tasks, 2) lower-level AI augments higher-level HI, and 3) at a given intelligence level, AI first augments, and then replaces HI. General principles are proposed to guide marketers, consumers, and researchers for collaborative intelligence in marketing.

This paper goes beyond the existing studies in that it lays out the multiple ways and conditions that AI and HI can collaborate, and delineates the collaboration scenarios for both marketers and consumers. Although the application of AI in marketing is gaining popularity in practice, existing studies tend to focus on algorithms for applying AI, the macroeconomic impact of AI (Huang, Rust, and Maksimovic 2019), the replacement effect of AI on human labor (Huang and Rust 2018), the benefits of AI at different stages of service process (Huang and Rust 2021a), or the strategic use of AI for marketing planning (Huang and Rust 2021b); work that guides the collaborative use of AI in marketing is limited.

In summary, this paper contributes to our understanding about the use of AI in marketing by providing a systematic framework for collaborative AI, which leverages the collaborative intelligence of AI, marketers, and consumers, and considers AI’s augmentation and replacement of HI from both the marketer’s and the consumer’s perspectives. We specify conditions for how marketers and consumers can actively leverage AI at different intelligence levels. With this big picture framework, that provides a deeper understanding of collaborative intelligence in marketing, marketing stakeholders can maximize the benefits of AI.