Generative AI for Visual Marketing Content Creation
(A2025-126167)

AUTHORS

Yannick Exner, Technical University of Munich; Tijmen Jansen, University of Hamburg; Maximilian Witte, University of Hamburg; Jochen Hartmann, Technical University of Munich

ABSTRACT

Generative AI is revolutionizing the marketing industry and research. This session curates a collection of working papers that study the intricate challenge of visual content creation with generative AI. These contributions offer practical and theoretical perspectives that are highly relevant to both industry professionals and academic researchers. The papers investigate digital advertising performance, the combination of AI with computer-generated imagery (CGI), the creation of influencer content that balances brand and creator interests, and the development of multi-modal marketing assets for music promotion. Each study approaches the technology from a distinct angle, collectively contributing to our understanding of how generative AI tools can be integrated into existing marketing processes for visual marketing content creation. The first paper, “AI in disguise — AI-generated ads outperform human-made ads if they don’t look like AI” by Yannick Exner, Jochen Hartmann, Oded Netzer, and Shunyuan Zhang examines how AI-generated advertising content performs compared to human-created ads through analysis of over 2 million real-world ad observations. It reveals that AI-generated ads can outperform human-made ads, but only if they do not look like AI. The second paper, “Visual Precision in Advertising: The Power of Parametric Image Control” by Tijmen Jansen, Claus Hegmann-Napp and Mark Heitmann, investigates the integration of generative AI and CGI to create optimized visual advertisements, showcasing how combining creativity of generative AI with precise control over visual parameters of CGI enhances engagement and advertising effectiveness. The third paper, “Automating fit in visual influencer marketing: Leveraging generative AI to align interests of marketing and influencers” by Julia Rosada, Maximilian Witte, Mark Heitmann, and Harald van Heerde studies whether generative AI can improve the creation of visual influencer content to align the objectives of influencers and sponsoring brands. Throughout five empirical studies they find that generative AI can create influencer content that outperforms actual sponsored influencer posts in engagement and mindset metrics. The fourth paper, “Marketing Music with Generative AI” by Benedikt Roder and Jochen Hartmann investigates how a custom generative AI pipeline can generate multi-modal marketing assets that combine audio, visual, and textual elements that enhance the real-world effectiveness and efficiency of social media campaigns for music artists. This session combines high-quality, practitioner-oriented research, offering insights into how generative AI can be leveraged to create visual content in marketing. These findings have the potential to guide marketers in developing more engaging and personalized content strategies that resonate with target audiences.