Prompt Chaining
Prompt chaining is a technique where the output of one LLM prompt is used as the input for the next, creating a sequence of connected calls that collectively accomplish a complex task no single prompt could reliably achieve.
Understanding Prompt Chaining
Single prompts have reliability limits. Asking one LLM call to simultaneously read email, identify urgency, extract tasks, draft replies, check calendar availability, and create a meeting invite is too much — the model is less accurate when juggling many steps at once. Prompt chaining breaks this into a sequence of focused prompts: Prompt 1 reads the email and classifies urgency → Prompt 2 extracts action items from urgent emails → Prompt 3 drafts replies for each action item → Prompt 4 checks calendar for scheduling suggestions. Each prompt does one thing well, and the chain achieves the complex goal reliably. Chaining also enables validation between steps. After each prompt, you can check the output before proceeding — verifying that email classification looks correct before drafting replies, or confirming task extraction before creating tasks in your project manager. Prompt chaining is related to but distinct from agent loops. Chains are predetermined sequences; agent loops are dynamic, with the model deciding at each step what to do next based on observations. Most real AI systems use both patterns.
How GAIA Uses Prompt Chaining
GAIA uses prompt chaining for predictable multi-step workflows like email triage (classify → extract → draft → act) and meeting prep (identify attendees → retrieve context → generate briefing). The chain structure ensures each step receives focused, high-quality attention rather than asking one prompt to do everything.
Related Concepts
Agent Loop
An agent loop is the iterative execution cycle of an AI agent in which it reasons about the current state, selects and executes an action (often a tool call), observes the result, and repeats until the task is complete or a stopping condition is reached.
Chain-of-Thought Reasoning
Chain-of-thought (CoT) reasoning is a prompting technique that instructs an AI model to articulate its intermediate reasoning steps before producing a final answer, significantly improving accuracy on complex multi-step problems.
Structured Output
Structured output is a technique that constrains an LLM to respond in a predefined format — typically JSON or XML — enabling reliable programmatic parsing of model responses rather than free-form text.
Prompt Engineering
Prompt engineering is the practice of designing and refining inputs to AI language models to reliably elicit desired outputs, shaping model behavior without modifying the underlying weights.
Agentic AI
Agentic AI describes artificial intelligence systems designed to operate autonomously, making decisions and executing multi-step tasks with minimal human oversight.
