Designing Stateful AI Copilots

This unit explores UX patterns and design strategies for stateful SaaS copilots, including interface approaches, model specialization, and persona-sensitive interactions. Students will evaluate which UX techniques are augmented or made obsolete by AI and prototype dedicated UI solutions that balance capability, continuity, and ethical risk.

Learning Objectives

• Analyze user workflows and data persistence requirements to define stateful behaviors and session management needs for a SaaS copilot
• Apply AI-augmented UX techniques and pattern choices (e.g., full-screen UIs, reference panels, conversation-state affordances) to create low-fidelity prototypes for two copilot design ideas
• Demonstrate integration choices by specifying when to use specialized fine-tuned models versus system-level context/state, and produce a rationale comparing trade-offs in accuracy, cost, and privacy
• Evaluate ethical, accessibility, and persona-safety implications (e.g., designing for vulnerable users) and recommend mitigations for harmful or misleading copilot behaviors