This chapter examines the evolving landscape of generative recommendation and planning systems, which harness the capabilities of large language models (LLMs) to generate content, user profiles, and multi-step recommendation plans. We begin by exploring key content generation tasks such as personalized descriptions, summaries, and conversational responses. We then outline strategies for evaluating these outputs through benchmark construction and task-specific metrics. We then turn to sequential planning, where LLMs support multi-turn dialogue and goal decomposition to enable proactive, context-aware recommendation. Lastly, we introduce two practical tutorials: one on personalized profile generation, and another on multi-step task planning with recommendations.