AI Application Scenarios

Many people's first impression of AI is a chatbot. Chatting with it is fun, but after a few conversations, they feel it's just so-so. That's because you treat it as a toy, not a tool.

The real value lies not in chatting with AI, but inusing AI to solve specific problems, such as writing emails, modifying code, organizing meeting minutes, and analyzing data — these are where AI can save you time.

The goal of this chapter is not to list all possible application scenarios, but to give you a thinking framework for finding AI use cases, and then let you experience how to put AI to work through a few typical scenarios.

AI's value lies not in what it can do, but in how much time it can save you. Before using AI each time, think clearly: What problem do I want to solve?

Writing and Content Creation

This is one of the most mature and easiest-to-use AI application scenarios. As long as text is involved, AI can almost always help.

Article Writing and Polishing

AI doesn't write a perfect article for you, but helps youquickly turn ideas into text。

You provide the viewpoints and structure, and AI handles organizing the language and enriching the details.

Here are some commonly used writing prompt templates:

Template 1: Article outline generation

我想写一篇关于 [主题] 的文章，面向 [目标读者]。

请帮我生成一个逻辑清晰的大纲，包含：
1. 引人入胜的开头
2. 3-5 个核心观点
3. 每个观点配一个简单的例子
4. 有说服力的结尾

主题：[具体主题]
目标读者：[比如：职场新人、技术管理者、创业者]

Template 2: Article polishing

请帮我润色以下文字，保持核心意思不变，但让它更 [风格]。

风格选项：更简洁、更专业、更口语化、更有感染力、更有逻辑

要求：
1. 保持原意，不要添加新信息
2. 让句子更通顺流畅
3. 去除冗余表达

需要润色的内容：
[你的文字]

Template 3: Style conversion

Please rewrite the following content into a version in [target style].

Marketing Copy Generation

Writing copy is AI's strong suit—it can quickly generate multiple versions to help you find the best expression.

Email and Report Writing

Working professionals write emails and reports every day; this is the scenario where AI can save the most time.

Email writing template

请帮我写一封 [邮件类型]。

收件人：[收件人身份，如：领导、客户、团队成员]
核心信息：[你想说什么，用 bullet point 列出]
期望行动：[希望对方做什么]
语气要求：[正式/友好/简洁/详细]

参考信息：
[背景信息，如：之前的对话、相关数据]

Let's take a real example:

Actual email prompt

请帮我写一封邮件。

收件人：项目组全体成员
核心信息：
- 原定于周五的项目评审会改到下周三下午 2 点
- 会议室不变，还是 302
- 请大家准备好 5 分钟的进度汇报
- 有特殊情况不能参加的，明天中午前告诉我

期望行动：收到请回复，确认能否参加
语气要求：正式但友好，不要太生硬

参考信息：这是 example 项目的第 3 次评审会

AI-generated content is only a draft; you need to adjust details based on the actual situation. But at least, you don't have to start from scratch.

Hands-on Example: From Idea to Finished Product

We'll use a complete example to demonstrate how to write a technical blog with AI:

Step 1: Generate an outline

我想写一篇关于"AI 如何帮助程序员提高效率"的文章，面向刚开始接触 AI 的开发者。

请帮我生成一个大纲，包含：
1. 开头用一个程序员日常痛点场景引入
2. 3 个核心应用场景（代码补全、Bug 调试、代码重构）
3. 每个场景配一个具体例子
4. 结尾给出实际的行动建议

Step 2: Write in sections

请帮我写开头部分，用一个程序员的日常痛点场景引入主题。

要求：
- 用具体的故事开头，不要空泛
- 场景要真实，让程序员有共鸣
- 200-300 字

参考场景：周一早上，程序员小王发现上周五写的代码有 Bug，改了一上午还没找到原因，这时他想起可以用 AI 辅助调试...

Step 3: Polish the whole piece

请帮我通读全文，做以下润色：
1. 让段落之间的过渡更自然
2. 把太技术的术语用更通俗的方式表达
3. 确保文章风格统一
4. 检查是否有错别字或不通顺的句子

全文内容：
[你的完整文章]

Code-Assisted Development

AI is changing how programmers work—not by making programmers unemployed, but by allowing them to focus on more valuable things.

Code Generation and Explanation

Writing repetitive code, looking up documentation, and writing boilerplate code—these are things AI does best.

Bug Debugging and Fixing

When encountering a bug, asking AI first has become a habit for many programmers.

Debugging Prompt Template

我遇到一个 Bug，请帮我分析可能的原因。

现象：[描述发生了什么，而不是你认为哪里错了]
期望结果：[应该发生什么]
实际结果：[实际发生了什么]

相关代码：
[相关的代码片段]

错误信息：
[完整的错误堆栈信息]

已经尝试过：
[你已经做过的排查]

Key principles:Tell the AI what happened, not what you think went wrong.。

If you directly say "I think it's a database problem," the AI may follow your line of thinking and miss the real cause.

Code Refactoring and Optimization

AI can help you turn "working code" into "good code".

Refactoring Prompt Template

请帮我重构这段代码，目标是：
1. 提高可读性
2. 减少重复
3. 更好的命名
4. 添加必要的错误处理

要求：
- 保持功能完全不变
- 给出重构说明：你改了什么，为什么这样改

原代码：
[你的代码]

Uses Suitable for Non-Programmers

You don't need to be a professional programmer to use AI to write simple code to improve efficiency.

For example:

This code demonstrates how to use AI to summarize documents.

Even if you don't understand programming, you can use a similar approach—tell AI your needs and have it write a simple script to handle repetitive tasks.

Data Analysis and Summarization

AI enables everyone to do simple data analysis without needing complex statistical knowledge.

Document Summary Extraction

This is one of the most practical features—throw long documents, long videos, and long recordings at AI and let it extract key information for you.

General Template for Document Summary

请帮我处理以下内容，生成一份摘要。

内容类型：会议记录 / 文章 / 视频字幕 / 邮件对话 / 培训材料

要求：
1. 先给一个 100 字以内的总览
2. 列出 3-7 个核心要点
3. 提取所有的行动项（谁、什么时候、做什么）
4. 列出关键数据和重要结论
5. 如果有待决定的问题，单独列出来

内容：
[你的内容]

If you need to process video or audio, you can first use a tool like Whisper to convert it to text, and then let AI handle it.

Data Interpretation and Visualization Suggestions

You don't need to be a data expert to see insights from data.

Automatic Generation of Meeting Minutes

A one-hour meeting takes two hours to organize minutes—AI can shorten that to 10 minutes.

Meeting Minutes Prompt Template

请帮我把以下会议内容整理成一份清晰的会议纪要。

会议基本信息：
- 会议名称：[比如：example 项目周会]
- 时间：[日期]
- 参会人：[名单]

要求格式：
1. 会议概述（一句话总结这次会开了什么）
2. 讨论要点（每个点用简短的话总结）
3. 行动项（表格形式：任务 | 负责人 | 截止时间）
4. 待决定事项（还没结论的问题）
5. 下次会议安排

会议内容：
[录音转写的文字 / 你的笔记]

Learning and Education

AI is the most patient teacher—it won't think you're stupid, won't get impatient, and you can ask it to explain as many times as you want.

Personalized Learning Assistant

The best way to learn isn't "watching videos" but "conversation"—ask questions immediately, correct misunderstandings right away.

Learning Assistant Prompt Template

我想学习 [主题]，我目前的水平是 [水平描述]。

请帮我制定一个学习计划：
1. 先评估一下我应该从哪里开始
2. 给一个 4 周的学习路径，每周有明确的目标
3. 推荐一些学习资源（不需要具体链接，说类型就行）
4. 每周给一个小练习，检验学习效果

我的背景：
- 相关经验：[比如：零基础 / 懂一点 Python / 用过类似工具]
- 学习目标：[比如：能做一个简单的网站 / 通过某个考试 / 工作需要]
- 每周学习时间：[比如：5 小时 / 10 小时]

请用苏格拉底式教学法——多问我问题，引导我思考，而不是直接给答案。

Concept Explanation and Q&A

When you encounter a concept you don't understand, have AI explain it to you in multiple ways.

Concept Explanation Prompt Template

请用几种不同的方式解释 [概念]。

要求：
1. 用大白话解释（不要用术语）
2. 用一个生活中的例子类比
3. 用一个简单的代码/数学例子演示
4. 解释它和 [相似概念] 的区别
5. 指出初学者最容易误解的地方

我目前的理解是：[你的理解，可写可不写]

For example, you can ask like this:

Please explain "what is machine learning" in several different ways.

Requirements:

1. Explain in plain language

2. Use analogies from daily life

3. Compare with "traditional programming"

4. Point out the most common misunderstandings for beginners

Language Learning Assistance

AI is a great companion for practicing foreign languages—it can converse with you, help correct your essays, and explain grammar.

Creativity and Design

AI cannot replace creative workers, but it can help them turn ideas into reality faster.

Brainstorming and Creative Inspiration

The nemesis of creativity is "a blank mind"—AI can help you quickly generate dozens of ideas, then you pick a few to explore in depth.

Brainstorming Prompt Template

我想为 [目标]  brainstorm 一些想法。

背景信息：
- 我们的目标是：[比如：给 example 网站想一个新 slogan / 设计一个线上活动 / 想几个视频选题]
- 目标受众：[谁]
- 风格偏好：[比如：有趣 / 专业 / 感人 / 有冲击力]
- 已有想法：[你已经想到的，可写可不写]
- 不要：[你不想要的方向]

请：
1. 先给 10 个常规但靠谱的想法
2. 再给 5 个疯狂但可能有趣的想法
3. 最后给 3 个混搭想法（把两个不相关的东西结合起来）

每个想法用一两句话描述即可，不用展开。

AI-Assisted Design Workflow

Even if you don't know how to use professional design software, you can still create decent designs with AI.

Typical workflow:

• 1. Describe what you want in text: A modern minimalist website homepage, with blue and white as the main colors, a top navigation bar, a Hero section in the middle for product introduction, three feature points below, and contact information at the bottom.

• 2. Let AI generate the design draft: Use tools like Midjourney, DALL-E to generate a few images.

• 3. Iterate and optimize: The layout of the first image is good, but the colors are too bright; make them darker and the font a bit larger.

• 4. Convert to code: Give the final image to AI and let it implement it with HTML/CSS.

Design description prompt template

我需要一个 [设计类型]，请帮我描述得更详细一些，方便 AI 生成。

我的初步想法：
- 用途：[比如：公众号封面图 / PPT 背景 / 网站 Banner]
- 风格：[比如：科技感 / 手绘风 / 极简 / 复古]
- 主色调：[颜色]
- 包含元素：[必须有的元素]
- 不要：[不想要的元素]
- 参考风格：[可选，比如：像苹果官网那种感觉]

请帮我把这些想法扩展成一份详细的设计描述，适合输入到 AI 设计工具中。

Story Creation and Scriptwriting

AI can help you quickly generate story outlines, character settings, and even complete first drafts.

Story creation prompt template

我想写一个 [故事类型] 的故事，请帮我构思。

类型：[比如：科幻 / 悬疑 / 爱情 / 职场]
长度：[比如：短篇小说 / 长篇小说 / 电影剧本]
核心设定：[比如：一个能看到别人剩余寿命的人 / 一家专门帮人实现愿望的店]
主角：[主角的基本情况，可简单可详细]
想表达的主题：[比如：选择的重要性 / 人与 AI 的关系]

请帮我生成：
1. 一个有起承转合的故事大纲
2. 3-5 个主要人物的设定（性格、背景、在故事中的作用）
3. 3 个可能的结局（喜剧、悲剧、开放结局）
4. 几个关键的冲突点

Customer Service and Q&A Systems

Using AI to answer customers' common questions is one of the easiest enterprise application scenarios to implement.

The Basic Logic of AI Customer Service

AI customer service doesn't need to "completely replace humans"; as long as it can answer 80% of common questions, it is already very valuable.

The simplest AI customer service workflow:

• 1. Collect frequently asked questions: Organize the questions that customer service is often asked.

• 2. Prepare standard answers: Write a clear and accurate answer for each question.

• 3. Let AI match the questions: When a user asks a question, AI finds the most similar question in the FAQ database and returns the corresponding answer.

• 4. Escalate complex issues to humans: If AI cannot find an answer or the user is not satisfied, automatically transfer to human customer service.

Customer service response prompt template

你是 example 公司的客服机器人。请根据以下信息回答用户的问题。

公司信息：
- 我们是做：[在线编程教育]
- 主要产品：[编程课程、实战项目、1对1 辅导]
- 工作时间：[周一到周五 9:00-18:00]
- 联系邮箱：[support@example.com]

常见问题与答案：
Q: 课程可以退款吗？
A: 开课后 7 天内可以无理由退款，超过 7 天可以申请延期。

Q: 课程有效期多久？
A: 课程购买后有效期 1 年，可以反复观看。

Q: 有试听课吗？
A: 有的，每门课都有前 3 节免费试听。

用户的问题：
[用户的问题]

要求：
1. 如果问题在常见问题里，用自己的话复述答案，不要直接复制
2. 如果问题不在常见问题里，礼貌地说不知道，并建议联系人工客服
3. 语气友好、专业，不要太机械
4. 回答要简洁，不要太长

Enterprise Knowledge Base Q&A

"Feed" the company's documents, manuals, and FAQs to AI; employees can directly ask AI when they have questions, without having to search for documents everywhere.

The technical term for this scenario is RAG (Retrieval-Augmented Generation) — AI first finds relevant documents in the knowledge base, then answers questions based on the document content.

How to Find Your Own AI Use Cases

After seeing so many use cases, you still might not know: where should I start?

Here's a simple method to help you find AI use cases that work for you.

Step 1: Record Your Week

Get a piece of paper or a document and record what you do every day this week.

No need to be too detailed, just jot it down simply:

Monday: Wrote two emails, revised a report three times, attended two meetings, and put together a spreadsheet.

Tuesday: Wrote code, debugged bugs, wrote unit tests, and discussed requirements with the product manager.

Wednesday: Interviewed three people, organized interview notes, wrote a weekly report, and replied to customer questions.

...

Step 2: Identify Three Types of Tasks

In your records, mark the following three types of tasks:

• Type 1: Tasks you hate doing—Things that give you a headache just thinking about them, but you have to do them anyway.

• Type 2: Tasks you do repeatedly—Things you do once a week/month, with similar processes each time.

• Type 3: Tasks you never have time for—Things you want to do but never get around to, like learning something new or organizing materials.

These three types of tasks are exactly where AI can help you the most.

Step 3: Start with the Smallest Thing

Don't start by thinking "I'll use AI to rebuild my workflow"—that's too complicated.

Find one thingthat's the smallest, clearest, and gives immediate resultsand start with that.

For example:

• Not "have AI write my report for me," but "have AI help me make my report draft read more smoothly."

• Not "have AI do data analysis for me," but "have AI explain what this Excel formula means."

• Not "have AI write code for me," but "have AI add comments to this piece of code."

Step 4: Build Your Prompt Library

Save the prompts you've tried and found effective—a simple note-taking app is all you need. Copy and paste them each time you use them, then fine-tune based on the specific situation.

Your prompt library is your personal toolbox—the more you use it, the more natural it feels.