Search Conferences

Type in any word, words or author name. This searchs through the abstract title, keywords and abstract text and authors. You may search all conferences or just select one conference.


 All Conferences
 EMAC 2019 Annual Conference
 EMAC 2020 Annual Conference
 EMAC 2020 Regional Conference
 EMAC 2021 Annual Conference
 EMAC 2021 Regional Conference
 EMAC 2022 Annual
 EMAC 2022 Regional Conference
 EMAC 2023 Annual
 EMAC 2023 Regional Conference
 EMAC 2024 Annual
 EMAC 2024 Regional Conference

EMAC 2024 Annual


Novel Applications of Generative AI in Marketing
(A2024-120021)

Published: May 28, 2024

AUTHORS

Arvind Rangasawamy, PennState Smeal College of Business; Bernd Skiera, Goethe-University Frankfurt, Germany; Michael Haenlein, ESCP Business School; Lukas Jürgensmeier, Goethe University Frankfurt; Sumon Chaudhuri, ESSEC Business School; Arnaud De Bruyn, ESSEC Business School; Wenlan Yu, The Smeal College of Business, Penn State University; Ning Zhong, The Smeal College of Business, Penn State University

ABSTRACT

Generative AI (GenAI) offers new opportunities for marketing academics to apply existing GenAI models, and/or adapt these foundational models for marketing applications and research. In this session, the presenters will discuss three applications of GenAI in marketing to highlight the opportunities and challenges associated with applying these technologies to address research issues in marketing. The discussant will summarize the common themes and insights across these applications. Lukas Jürgensmeier and Bernd Skiera, “Using Generative AI to Provide Scalable Feedback to Multimodal Exercises in Marketing Analytics.” Detailed feedback on exercises helps learners become proficient in marketing analytics. However, such feedback is labor-intensive and expensive for educators. This study introduces a web application leveraging GenAI to autonomously provide feedback on multimodal exercises requiring coding, statistics, and economic reasoning. We compare the application’s feedback with human expert feedback for 4,349 solutions. Using OpenAI’s GPT-4, our application provides almost unbiased evaluations, correlates very highly with (r = 0.94), and deviates only 6 % from human evaluations. While GPT-4 outperforms GPT-3.5 by 31 – 39 %, it costs over ten times more. Compared to AI models, human evaluators are one to two orders of magnitude more expensive and require at least 20 times the evaluation duration. Sumon Chaudhuri and Arnaud De Bruyn, “Bots Bargaining with Humans: Building AI Super-Bargainers with Algorithmic Anthropomorphization.” Bargaining is increasingly being automated by firms using AI. However, little is known about the psychological impact this would have on consumers. We train a bargaining AI within a Generative Adversarial Network (GAN) framework and task it with reaching superior economic outcomes while appearing “human” in doing so (a process we refer to as algorithmic anthropomorphization). We then experimentally compare it to two alternative bot specifications: a primitive bot that mimics human behavior and a purely economic-efficient bot. Our results suggest that (a) all bots perform poorly on subjective evaluations, (b) while superficial anthropomorphization helps portray a bot as a human, it does not improve subjective evaluations, and (c) algorithmic anthropomorphization offers the promise of a solution, albeit imperfect. Wenlan Yu, Ning Zhong, and Arvind Rangaswamy, “Title: Mining the “Mind of the Market” for New Product Ideas: A Prompted GenAI Model.” We develop and test an automated model to generate novel and relevant new product ideas that helps overcome “fixation” in the creative process of idea generation. We leverage big data (user-generated and firm-generated), graph-theoretic modeling, and machine learning to simulate human analogical reasoning on a large scale. The model captures the links between products, product uses/benefits, and product features to create an aggregated “mind of the market.” To generate new product ideas, we tap into this artificial mind to identify product-feature associations that have a high probability of occurrence but are not currently associated with a focal product. We test our model by generating new product ideas for four different mobile apps. We compare the ideas generated by our model with ideas generated directly by GPT-4. On average, the customers who participated in our study rank the ideas from our model as having much greater novelty than those generated by GPT-4.