Building an agentic workflow to break up and answer complex questions
In this video, we show you how to rebuild LlamaIndex's built-in sub-question query engine using our new workflows feature, including:
* Implement a complex, agentic workflow step-by-step in a Jupyter notebook
* Use our sub-question query engine to break down and answer complex queries
* Utilize our workflow visualization tools to understand the flow of your application
* Set up and use query engine tools with real-world data
* See the power of our ReAct agents in action, handling multiple sub-questions simultaneously
* Combine results from multiple sources to answer complex, multi-part questions
* Take advantage of LlamaIndex's versatility in handling large volumes of data across multiple documents
We demonstrate these capabilities using a real-world example, analyzing San Francisco's budget changes from 2016 to 2023. This video showcases how LlamaIndex can be used to create powerful, flexible applications for complex data analysis and question-answering tasks.
Notebook: https://colab.research.google.com/dri...