Prompt chaining is an advanced prompting technique used with large language models (LLMs) in which a complex task is broken down into a sequence of smaller prompts, where the output of one prompt is fed as input into the next​.Instead of asking an AI model to solve a complicated problem in one go, the process is guided through multiple steps: the model is first prompted with an initial query, produces an output, and that output is then included in a follow-up prompt to continue the process. By chaining prompts in this way, the model can handle multi-step reasoning or interactively refine its answers​. For example, one prompt might instruct an LLM to generate a high-level plan for a task, and the next prompt uses that plan to request detailed elaboration on each step. This method is essentially a form of prompt engineering that leverages the model’s ability to maintain context across turns. Prompt chaining is particularly useful for complex Q&A, calculations, or creative tasks where intermediate results help reach the final answer​. It enables piecewise problem solving: the model’s initial responses can be reviewed or moderated (by a human or another model) before proceeding, which can improve quality and controllability of the final output​. In sum, prompt chaining guides LLMs through a series of connected prompts, allowing them to tackle complicated tasks in stages and produce more coherent, accurate results.