Skip to main content
ModelTerms

Agents & Tools · intermediate

Workflow vs Agent (workflow, agentic vs deterministic)

A workflow is a deterministic pipeline where humans hard-code the LLM call sequence. An agent lets the LLM decide which steps to take. Anthropic's recommended default is workflow first, agent only when needed.

Explanation

Most production LLM systems are workflows, not agents — they look like "retrieve, then summarize, then classify" with the human choosing every step. Workflows are predictable, debuggable, cheap, and most problems don't need more.

Agents — where the model picks the next step from a tool registry, executes it, observes, and loops — unlock open-ended capability but at the cost of unpredictability, runaway loops, prompt injection risk from tool outputs, and an order of magnitude higher cost.

Anthropic's "Building effective agents" essay codifies the discipline: start with the simplest workflow that works; introduce agentic steps only where genuine flexibility is required; cap loops with budgets.

Examples

  • A workflow: incoming email → classify → route to template responder. Deterministic.
  • An agent: a coding assistant given "fix this bug" with file-read, file-edit, and shell tools — it decides what to do.

When to use workflow vs agent

Always reach for workflow first. Move to agent only when the next step genuinely cannot be predetermined.

Frequently asked

What is Workflow vs Agent?

A workflow is a deterministic pipeline where humans hard-code the LLM call sequence. An agent lets the LLM decide which steps to take. Anthropic's recommended default is workflow first, agent only when needed.

What is an example of workflow vs agent?

A workflow: incoming email → classify → route to template responder. Deterministic.

How is Workflow vs Agent related to Agent?

Workflow vs Agent and Agent are both agents & tools concepts. An AI agent is an LLM-driven system that decides which actions to take, executes them via tools, observes the results, and iterates until a goal is met.

When should I use workflow vs agent?

Always reach for workflow first. Move to agent only when the next step genuinely cannot be predetermined.

Is Workflow vs Agent considered intermediate?

Workflow vs Agent is generally considered intermediate-level material in the AI and LLM space.

AgentAgents & Tools

An AI agent is an LLM-driven system that decides which actions to take, executes them via tools, observes the results, and iterates until a goal is met.

Tool UseAgents & Tools

Tool use is when an LLM can call external functions — APIs, code interpreters, databases, web fetchers — and read their results. The mechanism that turns chat into action.

ReActAgents & Tools

ReAct is a prompting pattern that interleaves reasoning ("Thought:") with actions ("Action:") and observations ("Observation:"). It is the foundation of most tool-using agents.

Multi-AgentAgents & Tools

A multi-agent system uses several LLM agents that talk to each other — a manager and workers, a debate, a pipeline — instead of a single agent doing everything.

Agentic CodingAgents & Tools

Agentic coding is an LLM-driven workflow where the model reads code, plans changes, edits files, runs commands, and iterates against feedback — autonomously closing tasks rather than just suggesting code.

Side-by-side comparisons

Sources