Quick answer: Design autonomous AI agents — planning, tool use, reflection, error recovery.
AI Agent Design is the discipline of building autonomous systems that can break down complex problems, use tools strategically, and recover from errors independently. Instead of rigid workflows, agents perceive their environment, plan multi-step solutions, and adapt in real-time. Examples include: a customer support agent that routes queries to the right knowledge base, an investment research agent that gathers market data and synthesizes reports, or a code debugging agent that tests hypotheses and refines solutions. You're essentially designing systems that think and act, not just respond.