Welcome

Thank you for purchasing the MEAP (Manning Early Access Program) for Multi-Agent Systems with AutoGen. To get the most benefit from this book, you should have intermediate programming experience and expertise in at least one programming language, preferably Python. Additionally, a basic understanding of Generative AI and Large Language Models (LLMs) will be beneficial.

In early 2022, the core capabilities of generative AI models began showing incredible promise in solving various text generation tasks out of the box (question answering, summarization, and code generation, for example). It became clear that the potential for creating new types of applications addressing new tasks was immense. One of the initial projects I worked on at the time involved developing a tool that automatically generated high-quality visualizations and charts from data sources. The system operated as a pipeline that included data summarization, visualization goal generation, chart creation, and self-evaluation stages — with aspects of this technology now implemented in several Microsoft products. Since then, and driven by the popularity of tools such as LangChain, it has become standard practice to break down tasks into stages and build modules driven by generative models, targeting each stage.