Scholarlink

Generative AI has opened numerous possibilities for constructing intelligent systems and assistants. As the patterns become more established, there is a need for a disciplined approach to specify requirements for these applications. This chapter presents the beginnings of a methodical framework for business and design considerations for building an effective AI copilot. Our learnings are based on insights gleaned from 250 \(+\) workshops with customers and partners and our in-house experience constructing copilot services and applications. We cover several topics that must be considered: from articulating a clear vision and value proposition, dissecting user personas and their corresponding scenarios, to designing a user experience and interaction model that addresses the needs of the intended audience. We discuss copilot AI UX principles as well as maintenance and continuous improvement.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

So, You Want to Build a Copilot?

  • 章节

  • 发表日期：

    2025年

查看全文
  • Michal Furmakiewicz,
  • Chang Liu,
  • Angus Taylor,
  • Ilya Venger

摘要

Generative AI has opened numerous possibilities for constructing intelligent systems and assistants. As the patterns become more established, there is a need for a disciplined approach to specify requirements for these applications. This chapter presents the beginnings of a methodical framework for business and design considerations for building an effective AI copilot. Our learnings are based on insights gleaned from 250 \(+\) workshops with customers and partners and our in-house experience constructing copilot services and applications. We cover several topics that must be considered: from articulating a clear vision and value proposition, dissecting user personas and their corresponding scenarios, to designing a user experience and interaction model that addresses the needs of the intended audience. We discuss copilot AI UX principles as well as maintenance and continuous improvement.