Agentic workflow vs AI agent

Why they differ

In a workflow the developer's code decides the order of steps; in an AI agent the model decides for itself as it goes.

Agents & tools

A fixed chain of steps written in code, where a language model does some of the steps but does not choose the order.

Formal

A system in which large language model calls and tool calls follow paths set in advance by the developer - in a line, in branches chosen by rules, or side by side - rather than the model deciding its own next step as an AI agent does.

In plain English

Like a relay race on a fixed route - each runner is fast and skilled, but where they run and whom they hand the baton to was settled before the start.

In practice

At a water utility, each customer email passes a fixed chain - a model sorts it by topic, a prompt written for that topic drafts a reply, code checks the draft against the price rules, and a customer adviser sends it.

Why it matters

For most business tasks a fixed path is cheaper, easier to test and harder to lead astray than a free-acting agent, so it is often the better first choice.

AI agent

Language models

An AI system that does not just answer but acts - it plans steps and uses tools such as email, files or web search to reach a goal.

Formal

A system, usually built around a large language model, that is given a goal and tool access, then chooses and carries out actions in a loop - calling tools, reading results, deciding the next step - with little human input.

In plain English

Like an assistant given your keys and a to-do list rather than a chat window - useful, but only as safe as the keys you hand over.

In practice

A ministry tests an agent that arranges meetings - it reads officials' calendars, books rooms and sends invitations under its own account, and every action it takes is logged for review.

Why it matters

An agent that reads untrusted text can be tricked into acting against its owner, so the harm it can do is set by what it may reach and do, not by how well it chats.

Shared connections

Human-in-the-loop (HITL)Large language model (LLM)Multi-agent system Prompt