Using Knowledge Bases in Workflows

Lesson Overview

In this video, you'll learn how to leverage knowledge bases within AirOps workflows to enhance your content generation capabilities. Discover how to semantically search your knowledge base, pass the results to an LLM, and generate relevant, up-to-date content.

• 0:00: Introduction and recap of the previous video on creating knowledge bases
• 0:20: Demonstrating the use of knowledge bases in a simple workflow
• 1:03: Passing knowledge base search results to an LLM for content generation
• 1:26: Potential applications and benefits of using knowledge bases in workflows

Key Concepts

Knowledge Base Search Step

The Knowledge Base Search step in AirOps workflows allows you to semantically search for a query within your knowledge base. You can input a search phrase directly or dynamically generate it using previous workflow steps or user inputs. This step returns the search results in JSON format, which can be easily processed by LLMs in subsequent steps.

Integrating Knowledge Base Results with LLMs

LLMs are highly capable of reading and understanding JSON data. By passing the Knowledge Base Search step's output directly into an LLM step, you can provide the model with relevant, up-to-date information to generate content. This integration enables the LLM to incorporate the knowledge base results seamlessly into its output, enhancing the relevance and accuracy of the generated content.

Key Takeaways

1. Knowledge bases can be used within AirOps workflows to provide relevant, fresh information to LLM steps at runtime. This architecture allows for the generation of up-to-date, context-aware content.
2. The Knowledge Base Search step semantically searches for a query within your knowledge base, returning results in JSON format. These results can be passed directly into an LLM step for further processing.
3. LLMs can easily read and understand JSON data, making the integration of knowledge base results seamless. By providing the LLM with relevant information from the knowledge base, you can enhance the quality and relevance of the generated content.
4. Potential applications of using knowledge bases in workflows include fetching previously written content, finding internal links on your website, discovering customer stories and reviews, and injecting product information into e-commerce content.
5. The combination of knowledge bases and LLMs in AirOps workflows opens up a wide range of possibilities for generating high-quality, context-aware content. Experiment with this powerful architecture to unlock new content generation capabilities.