dahatake/GenerativeAI-Prompt-Sample-Japanese

What it solves

Many people use generative AI tools like ChatGPT or Copilot but get poor results because their prompts are vague or lack context. This project collects practical, Japanese-language prompt examples that show how to get much better outputs, especially for summarizing meetings, writing reports, and planning sprints. It also includes a cautionary note about handling sensitive information.

How it works

The repository provides a set of before-and-after prompt examples. Each "better" example shows you how to add reference material (meeting notes, transcripts) inside tags, specify the desired output format, ask for multiple alternatives, assign roles, and continue conversations for deeper analysis. Some examples also demonstrate how to resume a long output from a specific point. The project also points to an interactive learning site (hosted via GitHub Pages under the docs/ folder) where you can play with 2D/3D visualizations of Copilot to learn these techniques.

Who it’s for

This is for Japanese-speaking users who want to improve their prompt-writing skills for generative AI tools like Microsoft 365 Copilot or Azure OpenAI Service. It is especially useful for business users, project managers, developers, and anyone who needs to handle meeting notes, progress reports, or sprint planning more effectively.

Highlights

• Concise before-and-after prompt examples that are easy to copy and adapt
• Shows how to include reference materials (meeting transcripts, agendas) inside clearly delimited tags
• Demonstrates asking for multiple approaches with pros/cons/resources
• Includes tips for assigning roles and continuing multi-turn conversations
• Links to a free interactive learning site with 2D/3D visuals
• Provides a list of external resources for further prompt engineering study