1.6 Making it Intelligent
- Nick (Deactivated)
- Olivia Thomajan (Deactivated)
- Lydia Hou (Deactivated)
- Rajkiran Alla (Deactivated)
An intelligent conversational agent comprehends the context in the conversation and personalizes the results.
This document describes how to add intelligence to your Orbita-powered conversational agent.
Core Product Modules
The two modules you have to get familiar with are:
Experience Designer
Flow Studio
In Orbita, you have to use both these modules to build the best conversational experience.
Experience Designer
Experience Designer is a powerful, low code, graphical development environment. In the Experience Designer, you can handle Events and Actions, integrate APIs, customize design, create webpages and chatbots, and so on.
See, 2.2 Introduction to Experience Designer for more details.
Flow Studio
A module that offers an interface very similar to flow charts for creating logical conversational flows.
More information at 2.3 Introduction to Flow Studio.
Advanced integrations
Orbita’s extensibility means you can leverage 3rd party services to make your conversational experience smarter.
Some organizations might want to integrate with existing microservices they developed such as custom machine-learning models. The Experience Designer can be used to integrate and support them if they can be accessed via a web-based endpoint.
Realtime analysis of linguistic characteristics
There are numerous 3rd party APIs that can be leveraged to make it more intelligent. Sentiment analysis using 3rd party API, Tonal analysis through voice sdk, Language detection and real-time machine translation to name a few.
These services can provide the system with additional information or context that can be used to make your conversational agent more intelligent and influence the bot’s behavior.
You can integrate these services with the help of the Experience Designer > http request node.
Capturing data in Information Retrieval
The meaningfulness of the conversations that people have is drawn upon what was said previously in a conversation. For a conversational agent to mimic this quality of human dialogue, it must be able to store and retrieve information. Orbita platform namely the Experience Designer has a number of powerful capabilities that support data storage and retrieval.
For example, you might want to store user preference data to an object in the payload. You can define the type of fields that are to be used to store this information using a custom data schema. Then you can use the dynamic data manager node in the Experience designer to store and retrieve this information. See, How to retrieve partial conversation using flow studio
Adjusting how the virtual agent responds
Understanding is not the only part of making a conversational agent intelligent, it should respond intelligently too. If the voice agent and the chatbot say the same thing over and over again, it can feel very robotic.
Phrases
Orbita supports Phrases, also known as variations, that will introduce variety into the conversational dialogue.
For example, instead of greeting with a Hello in every conversation, the bot can say Hi, Hey, Hey there, Greetings, and so on. See, How do I create phrases for the voice assistant?
Multimodal content editor
The multimodal content editor can be used to change the pitch, rate, prosody, insert pauses, etc. See, What is the Multimodal Content Editor?
Text-To-Speech (TTS)
You can also change the gender and voice tone by leveraging third-party text to speech engines such as Google. See, How to configure a multi-lingual bot | Text to speech
Video Content
https://cdn.orbita.cloud/assets/orbita/general/makingitintelligent.mp4
Video Key Messages
Learn how to adjust how the virtual agent responds, perform information retrieval, and return information through accessing third-party services.
Capture data and trigger actions such as surveys, scheduling, FAQs, and learn advanced Integrations such as contextual awareness, multi-lingual translation.
Learn real-time analysis and response to linguistic characteristics.