AI & LLMs · Beginner
Workflow
Breaking a job into fixed steps and running them through in order automatically. Once the process is set, one sentence re-runs the whole thing instead of doing it by hand again.
TL;DRIn 30 seconds
- Break a job into fixed steps and run them through in order, automatically.
- Once the process is set, one sentence re-runs it instead of directing the AI from scratch.
- It follows a route decided in advance — meet something unplanned and it won't find its own way around.
OverviewIn detail
A workflow is a settled order for getting something done. Take "receive order, check stock, take payment, send notification, ship": what each step does and which comes first is fixed in advance, and the machine just follows. In an AI setting, it means having the AI walk a task from start to finish by fixed steps. Once the order is pinned down, a problem halfway through is easier to trace, because you know which stage to look at.
Directing an AI by hand every time gives unstable results and costs a lot of talking. Fixing the common processes means one click, or one sentence, re-runs the whole thing — less effort and more consistency. A lot of skills are exactly this: a particular workflow packaged up so it can be reused.
A content-production workflow might specify: read the topic, research it, write a first draft, self-check, then output. Trigger it once inside an AI tool and it runs down that line without you prodding it step by step. Kanban boards, approval chains and scheduled jobs all describe the same idea.
A workflow is about fixed steps; an agent is about working out the steps when it gets there. In real products the two are usually mixed, with the agent running inside one stage of the overall process and handling only the part that needs flexible judgement.
To judge whether a task suits a workflow, ask whether it is roughly the same every time and whether the steps can be pinned down in advance. The more fixed the steps and the more often you repeat it, the more a proper process pays off. Work that needs a fresh call every time is better handed to an agent.
MythsCommon misunderstandings
- ✕Thinking a workflow is the same as an agent ✓A workflow follows a set route; an agent finds its own route on the spot, and the two are usually used together
- ✕Thinking setting it up once is the end of it ✓When the inputs or the rules upstream change the process has to change too, or it just runs faster in the wrong direction
In practiceWhere you will meet it
- String the fixed steps of writing a piece into one process and run it with a click
- Automatically gather data and produce the weekly report on a set day
- Hand fixed actions like approvals and notifications to the system to run through
SummaryThe key points
- Splits a job into fixed steps and runs them in order
- Once fixed, one sentence re-runs the whole thing
- Steadier and less effort than directing it fresh each time
- Different from an agent, which works out its own steps on the spot
FAQ
What does Workflow mean?
Break a job into fixed steps and run them through in order, automatically.Once the process is set, one sentence re-runs it instead of directing the AI from scratch.It follows a route decided in advance — meet something unplanned and it won't find its own way around.
Where will I run into Workflow?
String the fixed steps of writing a piece into one process and run it with a click;Automatically gather data and produce the weekly report on a set day;Hand fixed actions like approvals and notifications to the system to run through。
What do people most often get wrong about Workflow?
Thinking a workflow is the same as an agent — but A workflow follows a set route; an agent finds its own route on the spot, and the two are usually used together;Thinking setting it up once is the end of it — but When the inputs or the rules upstream change the process has to change too, or it just runs faster in the wrong direction。
What are the key points about Workflow?
Splits a job into fixed steps and runs them in order;Once fixed, one sentence re-runs the whole thing;Steadier and less effort than directing it fresh each time;Different from an agent, which works out its own steps on the spot。
Common pitfalls It's a fixed route; an agent finds its own. The two are usually paired, not substitutes.
Also written asYou may also see it written as
These are all the same concept — the same thing, however it is written in a skill description. Do not mistake different spellings for different ideas.
Entry 7 of 186 · All entries
Keep readingRelated reading
AI Agent
An AI program that sets its own goal, breaks it into steps and calls tools to get the job done. You hand it a task and it works out how to do it — instead of only answering one question at a time.
LLM
A very large program that learned to predict the next piece of text after reading an ocean of it. Whatever you type, it is guessing what is most likely to come next — writing, translating and answering questions are all the same thing used differently.
嵌入(Embedding)
把一段文字换算成一串数字,让电脑能比较哪两句话意思更接近。搜索和知识库能在换了说法时也找对答案,靠的就是它。
Claude
A large language model built by Anthropic, known for understanding long documents, writing well and writing code. It comes both as a chat assistant on the web and phone, and in a dedicated form for programming.
Prompt
The message you send to an AI — the instruction or question you give it. How well that passage is written decides whether the answer that comes back is on target.
MCP
A standard interface that lets AI connect to outside tools and data in one uniform way. Build a tool to it once and any AI assistant can call it, saving you from writing a separate integration for each tool.
Where to go nextExplore more
Back to the skills themselves
Most of the concepts covered here come from skills that install into your AI client with a single command.