Not a Coder? Four AI Terms to Understand Before Your Team Starts Using Feishu

The features you use every day actually contain a practical beginner's introduction to AI

Want to learn AI but don't know where to start? The market is flooded with terms like prompt engineering, context processing, and Agent—each sounding more advanced than the last, as if programming knowledge is required just to get started. But perhaps you've already been using it in Feishu—you just didn't know its name.

This time, let's focus on one thing only: how to better leverage the AI features you've already encountered in Feishu, without learning to code. Even small changes to how you phrase requests can noticeably lift your day-to-day efficiency.

Markdown: The Universal Language of the AI Era

What exactly is Markdown? Simply put, it’s using simple symbols to add structure to text, making the structure easier for people and models to parse; the model output still needs review. For example, # indicates a heading, - denotes a list, and surrounding text with asterisks makes it bold. No need to click buttons or select fonts—formatting is achieved purely through input. Headings, lists and bold text can be represented in Markdown, but export behaviour should be checked in your Feishu version.

For instance, when you say, "I need a promotion plan for launching a new product, aiming to achieve 1,000 sales within 4 weeks with a budget under 50,000," AI interprets the structure as follows:

> # Task: New Product Launch Promotion Plan
> ## Goal
> Achieve 1,000 sales
> ## Constraints
> - Budget:
> - Under 50,000
> - Duration: 4 weeks
> ## Output Requirements
> 1. Divided into three sections: strategy, execution, and scheduling
> 2. Present budget in a table format

Understanding this opens up a new perspective in communicating with AI: structuring requests can improve clarity, but compare the output against your original task before relying on it.

How to Use It in Feishu?

Markdown Support in Feishu

📄 Document Export: Feishu Docs may support Markdown export—check your version and admin settings during a trial, then decide whether archived Markdown files suit your workflows

📁 Local File Conversion: If your Feishu version supports importing Markdown files, test whether headings, lists, tables and code blocks carry over as expected before relying on it

💬 File Preview in Messages: Depending on your Feishu version, Markdown files received in messages may be previewable in chat—confirm this behaviour during your trial

Prompts: Start by Providing Sufficient Context

Here’s a counterintuitive truth: in many cases, ordinary users don’t need to learn complex prompting techniques. Why? Because useful context often helps more than a bare instruction.

Giving AI your group chat history, meeting summaries, or past documents, along with a clear description of what you need, often works better than a carefully crafted prompt on its own.

Of course, you still need to state basic requirements. Try breaking your request into four parts:

• Background: What’s the situation? (“We’re holding a new product launch event next week”)
• Task: What exactly needs to be done? (“Help me draft a media invitation email”)
• Requirements: What are the standards? (“Formal yet engaging tone, under 200 words”)
• Reference: Provide a sample or style guide (“Follow the tone from our last event”)

How to Use It in Feishu?

Feishu chats, meetings and docs can provide part of the working context, subject to access permissions and data boundaries. Simply @Feishu AI within a document—it can draw on the document's content, so you avoid most copy-pasting. In practice, complete context usually beats a bare prompt.

HTML: An Alternative to PPT

Does making PPTs feel like this for you—spending half a day adjusting layouts, struggling with alignment until your eyes hurt, redoing everything when switching templates, and still worrying that the formatting will break on someone else’s computer after sending?

HTML web pages can be an alternative to PPT, depending on the audience and publishing requirements. Their advantages include:

Fewer layout issues: What you see is generally what others see, with fewer font, compatibility or version headaches than desktop slide files.

Link sharing: after hosting and access controls are configured, a page may be shared by link; opening behaviour depends on browser, login and administrator settings.

Broad device support: opens on phones, computers and tablets, and well-built pages adapt to different screen sizes.

How to Use It in Feishu?

Depending on your plan and admin settings, Feishu AI may be able to generate an HTML page shareable as a link—test this in your workspace trial, including whether external sharing is enabled.

Skill: Your Work Habits Can Be Summarized by AI

The first three are techniques for individual tasks, while Skills may help repeat a more consistent set of instructions.

What is a Skill? Put simply, it packages recurring instructions for reuse; actual behaviour should be verified in the current product version.

Take writing a report as an example. The first time, you might give AI detailed instructions on format, tone, and data presentation. Once these experiences are saved as a Skill, the next time you only need to say, “Follow the same format as last time,” and it can produce a draft quickly—consistency usually improves once your preferred format is captured.

Skills can help turn parts of a workflow into reusable instructions, with human review before delivery.

How to Use It in Feishu?

Feishu AI’s skill library may include ready-to-use Skills—browse what is available in your version and test one on a real task before adopting it.

If your plan supports it, you may be able to save a satisfactory AI conversation workflow as a custom Skill for reuse—verify whether custom Skills work as expected in your trial.

These four concepts can be combined in some workflows:

Use Markdown to structure a request → add relevant context → use HTML where suitable → package recurring instructions as a Skill. Test first with non-sensitive material and review the output.

Follow us to discover more AI application methods 👇

Explore previous content below