Agentic Workflow
An agentic workflow is a multi-step AI process where a model plans, takes actions through tools, observes results, and iterates toward a goal — replacing single-shot prompting with a loop that reasons, acts, and self-corrects.
Full definition
Agentic workflows decompose tasks into sub-steps (plan), call tools (search, code execution, APIs), incorporate results into context, and decide the next action. Patterns include ReAct (reason+act), reflection, planner-executor splits, and multi-agent orchestration. Andrew Ng's 2024 essays argued that agentic workflows on weaker models often outperform single prompts on stronger ones — making workflow design a first-class lever, not a model-selection afterthought.
Why it matters
Most enterprise AI value above the chat-interface layer is produced by agentic workflows: research, code generation, customer-service triage, compliance review. Teams that treat agentic design as a discipline ship qualitatively different products than those that stop at "ChatGPT for X."
Example
A market-research agent reads a brief, generates 12 sub-queries, runs web searches, downloads PDFs, extracts tables, cross-checks numbers, and produces a sourced 4-page memo — a 6-hour analyst task in 8 minutes.
Related terms
- AI Agent / Agentic AIAn AI agent is an LLM-driven system that, given a goal, plans a sequence of steps, calls tools or APIs, observes the results, and iterates until the goal is reached — going beyond single-turn chat to multi-step autonomous action.
- ReAct PatternReAct (Reasoning + Acting) is a prompting pattern that interleaves reasoning steps ("Thought") with tool actions ("Action") and tool results ("Observation"), letting an LLM iteratively decompose a task, query the world, and self-correct.
- Multi-Agent SystemA multi-agent system orchestrates multiple specialized AI agents — often LLM-powered, with distinct roles, tools, and prompts — that communicate, coordinate, and divide labor to solve a problem larger or more diverse than any single agent could handle reliably.
- Tool Use (LLM)LLM tool use (also called function calling) is the ability of a language model to invoke external functions, APIs, databases, or systems based on the user's request — turning the LLM from a text generator into a controller that can act on the world.
Source & further reading
Primary source: Ng, A. — "Agentic Design Patterns" (DeepLearning.AI) (2024).
Citation policy: this entry is part of the AIDOLS AI Implementation Glossary and may be quoted for research, journalism, and education with attribution to aidolsgroup.com/en/glossary/agentic-workflow/.