LangChain introduces a new type of agent executor called "Plan-and-Execute" agents, contrasting with previous "Action" agents. Inspired by BabyAGI and the Plan-and-Solve paper, these agents are designed for more complex long-term planning at the cost of increased calls to language models. The initial version is placed in the experimental module due to expected rapid changes. Plan-and-Execute agents separate higher-level planning from shorter-term execution, using a planner and an executor. Future directions include better support for long sequences of steps, revisiting plans, evaluation methods, and selection of execution chains.