This section describes the mechanism for customizing the behavior of your Digital Person using Behavior Tags, and how to use them on different versions of Human OS.
Behavior Tags are a mechanism for training your Digital Person to act out certain behaviors while they are delivering content. They are helpful because they trigger a behavior any time a word or phrase of your choosing is uttered by your Digital Person, throughout your conversation corpus.
Behavior Tags enable you to add a limited set of facial expressions and gestures to selected words, phrases, and sentences in your digital person's speech.
If you’re working with an English-speaking digital person on Human OS 2.0+, or a Japanese-speaking digital person with Human OS 2.3+, you can take advantage of the Real-Time Gesturing feature, and use it in conjunction with Behavior Tags.
If you’re working with a digital person running on Human OS 1.7, or your digital person does not speak a language that supports Real-Time Gesturing, you are still able to use the full potential of Behavior Tags.
Real Time Gesturing is the real-time autonomous animation of a digital person, based on their speech. It’s a superpower for the conversation designer: the words they write autonomously bring the digital person’s behavior to life. Appropriate expressions and gestures are performed as the digital person speaks. This behavior will match the meaning of the speech but be slightly different every time, reflecting real human behavior.
With Real Time Gesturing Enabled, Behavior Tags become a way to add additional emphasis to your conversation based on words and phrases associated with your use-case, your brand, or to add emotional performance where there is subtextual meaning known to the author of the conversation.
You can use the behavior tags to highlight words and phrases relevant to your use case. For example, you might want to add brand and product-related words to Smiling Gesture (Word-Based).
You can use the behavior tags to override the autonomous behavior if it’s not quite appropriate for your use case. For example, if you want the digital person to appear sad but their dialogue is not sad, you could add it to the Concerned Expression (Sentence-Based).
You can also use Neutral Long (Sentence-Based) and Neutral Short (Word-Based) to define sentences and words that will temporarily deactivate the real-time gesturing entirely.
This is for digital people on Human OS 1.7, and digital people without a language that supports Real-TIme Gesturing.
With Behavior Tags, you can fine-tune the head and neck gestures and emotional behavior of your Digital Person.
Behavior Tags enable you to add a limited set of facial expressions and gestures to selected words, phrases and sentences in your digital person's speech.
Adding expressions and gestures like a specific eye gaze can really help convey the inner life of your Digital People and in so doing, will enhance the relatability and naturalness of the experience for the end-user.
We recommend taking a thorough approach to adding Behavior Tags to your corpus, as this will add liveliness to your digital person’s behavior throughout the conversation, and emotionally appropriate expressions to their speech.
You can use the behavior tags to highlight words and phrases relevant to your use case. For example, you might want to add brand and product-related words to Smiling Gesture (Word-Based).
You can use the behavior tags to highlight the emotional meaning of your digital person’s dialogue. For example, if you want the digital person to appear sad, you could add phrases or whole sentences to the Concerned Expression (Sentence-Based).
There are two types of Behavior Tags: sentence and word-based tags.
Sentence-based Tags
These expressions are designed to generate an overall emotional feeling across the entire length of a sentence. Because they last for a relatively long time, they are subtle and are composed of a blend of many different facial gestures.
Word-based Tags
These expressions are more pronounced and should be treated like punctuation for emotional effect. Short expressions typically extend across the length of a single word or very short phrase.