Data, Models, and Experimentation

This unit explores data strategies and AI model families for behavioral applications, and teaches how to design experiments that produce actionable insights. Students learn to define minimum viable datasets, compare classical and modern AI approaches, and apply segmentation and intent-prediction methods to map customer journeys—building on foundational concepts and preparing for applied implementation and evaluation in subsequent units.

Learning Objectives

• Analyze trade-offs between data volume, data quality, synthetic and survey sources to define a minimum viable dataset (MVD) and justify data collection priorities for a behavioral AI problem
• Apply experimental design principles (e.g., control groups, randomization, metrics) to plan and simulate an A/B or multivariate test that evaluates a behavioral AI intervention
• Differentiate between classical AI, machine learning, deep learning, generative AI, and biologically inspired models by characterizing their data requirements, interpretability, and suitability, and recommend the most appropriate model class for a given business scenario
• Evaluate segmentation strategies using behavioral and intent signals to predict customer intent and map customer journeys, producing a prioritized data-collection and modeling roadmap for deployment