How to Build a Six Thinking Hats AI Agent: From Concept to Deployment

Background

After attending a talk on Edward de Bono's Six Thinking Hats, the author realized the tool’s potential for systematic, multi‑angle thinking but noted its high usage cost in daily life. With the rise of AI, they envisioned an inexpensive AI assistant that could apply the hats automatically.

Six Thinking Hats Overview

The Six Thinking Hats, created by Edward de Bono, assign a specific perspective to each colored hat:

White – objective facts and data.

Red – emotions and intuition.

Black – critical judgment and risk identification.

Yellow – optimism and opportunities.

Green – creative ideas and solutions.

Blue – organization and control of the thinking process.

Using the hats sequentially enables comprehensive, balanced decision‑making.

Benefits

Structured thinking, full‑scope problem analysis, higher decision quality, improved team collaboration, reduced emotional bias, encouragement of innovation, better communication, and flexible application across personal and organizational contexts.

Agent Introduction

Recent articles differentiate between Chatbot, Copilot, Agent, and Intelligence. Agent‑level AI can handle most tasks, leaving humans to set goals, provide resources, and supervise results. Platforms like AutoGPT showcase this capability, approaching AGI.

Agent Platforms

Many platforms exist, but most still resemble chatbots rather than true agents. Examples include:

Tongyi Qianwen – supports creative studio, templates, and free‑form creation with advanced settings for greetings, inspiration, skills, and knowledge upload.

Dify – offers a production‑ready solution similar to LangChain, open‑source deployment, data flexibility, and security.

Agent Creation Process (Tongyi Qianwen Example)

The author demonstrates building a Six Thinking Hats agent, covering avatar selection, prompt design, and iterative debugging.

Prompt Example (Version 1.0)

## 技能
请根据“六顶思考帽”相关知识，分别用 6 顶思考帽对问题进行分析和提供建议。

Initial results were acceptable but showed inconsistencies such as missing hat explanations and occasional English output.

Prompt Improvements (Version 2.0)

## 技能
请根据“六顶思考帽”相关知识，分别用 6 顶思考帽对问题进行分析和提供建议，给出每个方面的答案时，先简要介绍一下每顶帽子的作用。</code><code>## 示例
用户输入：康威定律
模型回答：
"""
康威定律（Conway's Law）是由计算机科学家梅尔文·康威在 1967 年提出的…
白色帽子（信息帽）作用：<该帽子的作用> 分析：<对应的分析>
…（后续每顶帽子同样格式）
"""
## 要求
- 务必使用中文回答

These changes enforce a brief description of each hat’s role, require Chinese responses, and specify output format, resulting in more stable and higher‑quality answers.

Testing and Deployment

After satisfactory testing, the agent can be shared via a generated link. Continuous prompt refinement is essential for complex or highly personalized scenarios.

Conclusion

The guide demonstrates how the Six Thinking Hats method can be transformed into an AI agent, illustrating the full workflow from concept to implementation and emphasizing the importance of prompt engineering, iterative debugging, and balanced AI usage.

Easter Egg

The author also created auxiliary assistants for title generation, article polishing, critique, image selection, and mind‑map conversion, all of which helped produce this article.