Guide to invoking functions within the knowledge base of your voice agent
Using functions, you can enhance the knowledge base of your voice agent by including custom functionalities. Here’s how you can invoke a function in the knowledge base:
Now that we’ve ensured our function works, we want to get our voice agent to use it. Open the knowledge base and start writing your instructions.
You can add a function into the knowledge base in three ways:
All of these methods will bring up the function context menu where you can search through all your agent functions and select the one you want. From this menu, you can also create a function in context and populate its definition later.
You can use a function as many times as you like to cover all the use cases you want your voice agent to handle.
It may require some iterative testing to get your voice agent to call your function and behave as expected.
The example below has been found to work in many cases, but you will likely find better ways of constructing your conversation through trial and error.
Remember, we passed a the JSON location object as a string to the LLM.
By following these steps, you can successfully invoke functions in the knowledge base of your voice agent.
For any questions or further assistance, please refer to our Support Page or contact our support team.