Historically, punctuation characters were removed from user utterances in voice channels when using language models in Zammo.
This approach made it easier for conversation designers to use regex patterns or define trigger phrases—even when those phrases included punctuation marks.
It can be useful in certain circumstances, however, to have greater control over how punctuation characters are handled in the language model that underpins your conversational AI solution.
The setting Enable Automatic Punctuation
makes it possible for you to enable or disable the removal of trailing punctuation characters in your language model.
In this guide, you’ll learn how to manage punctuation in your conversational AI application.
<aside> 💡 By default, this is enabled and punctuation is passed by default by the various channels. You can opt to disable it to remove the lagging punctuation from all channels.
</aside>
The table below shows the output of the new setting with the sentence: This is a new sentence
Setting | Example Output |
---|---|
Enable Automatic Punctuation = Off | This is a new sentence |
Enable Automatic Punctuation = On | This is a new sentence. |
This feature can be accessed directly in UI Builder or from the Zammo Portal for any business that you manage.