Learning Timeline
Key Insights
Context Quality Over Basic Prompting
A common mistake among AI users is only providing basic information. For high-quality results, you need to provide dozens of pages documenting your vision and work processes.
The 'Data Soaking' Technique
Don't rush to give tasks to the AI. Let it 'soak' in your entire life history, company documents, and database first so it develops a robust understanding of your context.
The Importance of Data Organization
AI performs significantly better when provided with a well-organized 'Research Database' compared to a pile of unstructured data.
Prompts
Storytelling Style Analysis
Target:
ChatGPT / Claude / OpenClaw
Compare the data in the research database versus the actual final output in these transcripts. Analyze how I weave facts into a storyline, what specific examples I provide, and the level of depth in my communication. Use this analysis to replicate my personal style in future tasks.
Step by Step
Context Engineering Process for AI Agents
- Prepare at least 12 Markdown files containing your personal profile, life vision, and in-depth work details.
- Create separate Markdown files explaining your company’s organizational structure, operational departments, and long-term business vision.
- Document every specific work process (SOP) in a clear Markdown format for the AI's reference.
- Organize a structured research database and feed it into the AI system as factual reference material.
- Extract and upload transcripts from YouTube videos or existing public media to teach the AI your specific tone of voice.
- Allow the AI a 'soaking' period to absorb all the data before assigning its first task.
- Instruct the AI to analyze the differences between raw data (research database) and the final output (videos/articles) to understand your storytelling techniques.
