Generative artificial intelligence is reshaping marketing practice, yet marketing education lacks clarity on which forms of human expertise remain distinctive and assessable when generative systems accelerate production. This study develops a practice-grounded framework of AI-resilient marketing capability through a qualitative investigation of 33 in-depth interviews with practitioners across sectors and levels of AI integration. Rather than testing predefined competency models, the study inductively surfaces how practitioners identify non-delegable human contribution in AI-mediated workflows. Analysis yields the Voice–Judgement–Taste (VJT) framework: voice as expressive authorship that conveys identity and audience relationship; judgement as deciding when AI outputs can be trusted, adapted, or rejected; and taste as evaluating which outputs are worth developing and ensuring coherence across them. Together, these interdependent capabilities explain how professionals manage productivity–authenticity tensions in AI-assisted work. Findings show that human value shifts from producing content to shaping it—for example, revising tone and identity, verifying accuracy and risk, and selecting the most appropriate option from multiple viable outputs. By translating broad graduate attributes into observable practices, the study advances conceptual clarity in marketing education and offers scalable assessment principles for making human contribution visible in AI-rich learning environments.
Gonsalves et al. (Wed,) studied this question.