深习堂
课程价格指南博客

开放课程Skill DevelopmentAI Courses

Practical Prompt Templates

Choose a prompt structure that gives AI less room to guess what you need

开始学习
全部章节
6
Locale
English
课程类型
开放课程

Use four prompt structures to explain a task, provide context, define sources, and set requirements. Choose a structure that addresses the main information gap, identify the most important problem in the first result, and revise the relevant part. Compare both results to judge whether the change helped.

适合人群

  • People who already use AI but repeatedly have to clarify what they need
  • People who have seen prompt templates but are unsure which one to use
  • Professionals who use AI for writing, summarizing, organizing information, and office tasks

解决的问题

  • Requests such as “write this” or “improve this” leave AI guessing about your purpose and priorities
  • You fill in the same long template for every task, but the results do not improve
  • After a poor first result, you change so much that you cannot tell which revision helped

学完收获

  • Identify what AI has to guess and which missing detail matters most to the result
  • Explain what each structure helps clarify and which sections a particular task can do without
  • Choose a suitable structure and clarify the task, source material, or success criteria
  • Find a specific problem in the first result, revise the relevant part of the prompt, and compare both versions

课程亮点

  • Compare ways of organizing prompts from OpenAI, Anthropic, Microsoft, and Google
  • Compare all four structures through the same quarterly-report scenario
  • Each structure includes a diagram, template, worked example, and one revision
  • Finish with your own prompt and both the initial and revised results

全部章节

6 个章节

Lesson 1: Find the Gaps That Affect AI's Output

Identify the information AI must guess from a vague request, and judge which missing piece is most likely to change the result.

Lesson 2: Explain the Work Clearly with the OpenAI Approach

Write a lightweight work handoff using task, helpful context, and ideal output, and judge from information complexity whether to expand it into an enterprise structure.

Lesson 3: Make Your Standards Visible with the Anthropic Approach

Choose clear instructions, reasons, examples, or success criteria based on how a first output deviates, and use one cue to revise the prompt.

Lesson 4: Manage Sources and Deliverables with the Microsoft Approach

Organize a source-dependent task with Goal, Context, Source, and Expectations, and write clear rules for source scope, authority order, and gap handling.

Lesson 5: Turn Keywords into a Generation Brief with the Google Approach

Expand scattered keywords into a runnable generation Brief using Persona, Task, Context, and Format, and revise one part based on output deviation.

Lesson 6: Pick One Structure and Complete a Real Interaction

Choose a suitable starting structure for a real task, make a single-point revision against one observable deviation, and judge from two versions of the result whether the improvement worked.