Creating with algorithms: how advertising agencies integrate, differentiate, and transform through generative AI

Abstract

Generative artificial intelligence (GenAI) is rapidly transforming creative industries, yet we lack understanding of how creative organisations integrate these technologies, sustain distinctiveness when working through shared algorithmic systems, and transform their capabilities under technological disruption. This thesis addresses these gaps by examining how advertising and creative professionals navigate the tensions between AI-enabled efficiency and the distinctiveness that creative markets reward. This thesis investigates how generative AI transforms creative work across multiple levels of analysis. Adopting a qualitative, inductive approach using Gioia methodology, three complementary studies examine this phenomenon in the Chinese advertising and creative industry. Study 1 employs semi-structured interviews with 20 advertising practitioners to examine the AI integration process. Study 2 conducts 13 months of embedded ethnography in an advertising agency, comprising 720 hours of participant observation and 32 interviews, to investigate how creative teams resist algorithmic homogenisation. Study 3 undertakes a multiple-case study of four creative firms, drawing on 27 interviews, to trace how organisational creative capabilities transform. The thesis develops three interconnected theoretical contributions. Study 1 introduces a generative AI-integrated co-creative process model comprising four stages: readiness, co-creativity, validation, and execution, demonstrating how creativity becomes distributed between humans and AI through iterative collaboration. Study 2 theorises legitimacy-embedded algorithmic distinctiveness work as a coordinated practice system through which creative professionals resist convergence, comprising inscribing, curating, marking, and enforcing practices. It further identifies the allocation regime as the mechanism that helps account for divergent outcomes, including an efficiency-visibility trap through which visible acceleration can trigger devaluation pressure. Study 3 specifies two aggregation mechanisms linking individual experimentation to organisational capabilities: Performance Revelation and Institutional Codification. These mechanisms underpin the emergence of five transformed capabilities for AI-augmented creative work: Hybrid Idea Orchestration, AI-Mediated Socialisation, Creative Digital Infrastructure, Algorithmic Agility, and Platformised Openness. The thesis makes theoretical contributions to creativity and creative process theory, algorithmic organising, professional work and expertise, and dynamic capabilities research. Practically, it offers guidance for creative agencies navigating AI integration, clients evaluating AI-augmented creative services, and AI tool developers designing for professional creative contexts. Collectively, these studies reveal that sustaining creative value in an algorithmic age requires coordinated organisational effort rather than individual prompting skill alone.

Date of Award	19 Jul 2026
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Russa Yuan (Supervisor) & Martin Liu (Supervisor)