This taxonomy of marketing functions can guide implementation of artificial intelligence from machine learning to generative AI. Enablement proceeds from Interpretation (research and analytics) to Direction (strategy and tactics) to Optimization. Engagement involves Creation of personalized products and services that drive customer Connection. AI implementation can proceed in either direction.
Develops a taxonomy of marketing functions in five domains and maps relationships to support alternative pathways and a scorecard for AI adoption.
Unsold perishable food waste costs retailers and generates greenhouse gases. New stores face a “cold start” problem of what and how much to stock but can reduce waste by nearly two-thirds within two years. Regulators and investors need to allow for some waste as part of this dynamic learning process.
Analyzes data from a Japanese grocery chain on perishable food product ordering, sales and waste to determine whether—and how quickly—new stores can learn to match local demand and reduce waste.
Livestreaming by professionals such as physicians and financial advisors can generate demand. However, unlike social influencers who use aspirational or entertaining information and discounts to promote third-party brands, professionals must rely on formal credentials, demonstrated expertise, audience interaction, substantive and concrete content, simple language, and a neutral, professional emotional tone.
Analyzes consultation data for over 7000 physicians on a major Chinese healthcare platform to determine the features of livestreaming that impact demand.
While generating ads with LLMs/AI models is cheap and efficient, evaluating consumer response can be expensive. Iterative prompt-optimization-based self-improving AI incorporating Bayesian optimization can prompt high-performing creative using the fewest evaluations possible. Resulting ads can be validated against target audience preferences derived from a “digital twin” database.
Develops and validates an iterative prompt-optimization-based self-improving AI that uses Bayesian Optimization in the language space to generate prompts for high-performing ads using the fewest evaluations possible.