Kouzi Develops AI Agent Applications

Learn how to develop AI agent applications through hands-on projects focused on the Kouzhi platform’s workflows, knowledge base, prompts, and database.

Bibliographic Information

• Book Title: Button DevelopmentAI AgentAgent Applications
• Authors: Song Lihuan, Wang Dongjian, Chen Mingyi, Cheng Dongsheng
• Publisher: Tsinghua University Press
• Publication Date: June 1, 2025
• ISBN: 9787302692188
• Language: Simplified Chinese
• Publisher's suggested retail price: 99 RMB; non-real-time transaction price

What is this book about?

To turn a conversational AI into an application capable of handling tasks step by step, the key is to link together inputs, models, tools, and outputs. Using the Koupi platform as a practical environment, this book focuses on agent configuration and workflow setup to help readers break down business ideas into tasks that can be validated step by step.

Author and Background of the Work

This book was co-authored by Song Lihuan, Wang Dongjian, Chen Mingyi, and Cheng Dongsheng. According to the publisher’s description, the authors bring diverse backgrounds in technical architecture, enterprise services, practical applications of Kouzi, and artificial intelligence education; consequently, the content focuses more on platform applications and teaching practices.

Content and Structure

The book consists of 16 chapters. The first half covers concepts of intelligent agents, an introduction to the platform, plugins and cards, workflows, knowledge bases, prompts, and databases; the second half explores various scenarios, including content generation, language learning, interior design, marketing posters, customer service, PowerPoint presentations, news videos, and travel assistants. The publisher also provides supplementary materials, including a prompt file, course materials, accompanying illustrations, and 190 minutes of instructional videos.

Who Is This Book For?

This book is ideal for operations professionals, product managers, educators, and beginners who want to build their own AI applications, as well as for teams that wish to validate business processes before diving into code development. Readers with prior development experience may want to focus on the sections covering the knowledge base, node orchestration, and multi-agent collaboration to assess the limitations of low-code solutions.

Suggested Pronunciation

Start by building a minimal workflow: receive input, call the model, and check the output. Next, select a real-world but low-risk scenario, and then add a knowledge base or database. For each new node added, test normal inputs, missing information, and failure scenarios. Enterprise data should be anonymized first, and you should confirm the data processing requirements of the platform and plugins.

Value and Limitations

Its advantage lies in translating the abstract concept of an “agent” into concrete platform operations. Its limitations, however, stem from platform dependencies: since this book was published in 2025, page layouts, plugins, models, and billing rules may have changed, so you’ll need to refer to Koushi’s latest documentation. The fact that an example runs successfully does not mean it meets production standards; before deployment, you must still perform checks on permissions, costs, error handling, and output quality.

Resources and Purchasing Guide

Bibliographic Information and Basis for Content StructureTsinghua University Press Official Book Pageup toOfficial CatalogCompiled and verified on October 18, 2026. Reading recommendations and analysis of limitations are based on the judgment of this site’s editors. Purchase options can be viewed on the publisher’s page; actual inventory and final sale prices are subject to the retailer’s listing.