Create a knowledge base
Step one: Install your OpenAI connector
OpenAI powers a search functionality in chatbots. By installing OpenAI, your chatbot will be able to search through your knowledge base content and return results much more efficiently.
In Bird, go to Connectors and follow the instructions to install OpenAI.
Step two: Create a new knowledge base and enable OpenAI
Go to Knowledge bases.
Click Add new.
Enable the OpenAI connector that you installed in step one.
Step three: Create a folder structure
The documents in your knowledge base need to be organized into topic-based folders.
Nest folders within each other, starting with broader topics as the top-level folders, and nesting narrower sub-folders within them. This is particularly important if you have content in multiple languages.
Give your folders distinct, recognizable names. This will help you and your AI chatbots to find information easily.
Click Add new > Folder.
Enter a name for your folder.
Click Get started.
Repeat the process by clicking Add new in the top right hand corner, and creating more folders.
To create a folder within a folder, select one of your newly created folders. then click Add new > Folder.
While you can modify the structure of your knowledge base later, it's best practice to plan for scalability and future additions.
Step four: Add your documents
Now that your file structure is set up, it's time to add your documents.
Select the folder where you want to add your document.
Click Add new > Document.
Enter a unique name for the document.
In the content editor, enter your text content.
Use the text styling tools to add headings, bullets, or other styles to your document.
If your document is a work in progress, click Save as draft. It won't be accessible by teammates or chatbots.
If your document is ready, click Publish to make it available for teammates and chatbots.
Repeat this process until you have added all of your knowledge base articles.
Step five: Integrate your knowledge base with an LLM chatbot
Now that your knowledge base has been set up, you can deploy an LLM chatbot to handle customer conversations using content from your knowledge base.
Last updated