Jacada's Intelligent Assistant (IA) is an interactive virtual agent that creates conversation-like interactions with your users, and helps reduce the resources that are needed to enhance their service experience.

IA chatbots use the power of Natural Language Processing (NLP) to understand the intentions of a user and translate these into actionable transactions in your flows.

The IA conversational approach enables an Interact user to enter free text in their native language to describe what they need or want to do. For each user entry, the IA analyzes the text and responds according to its knowledge of the user, their intention, and your service or business needs.

For a Designer's perspective on how Jacada's Intelligent Assistant works, see What Is the Intelligent Assistant?.

Note: The IA needs to be set up specifically for each Interact account. Contact your Jacada Support representative to get started.

Components of the Intelligent Assistant

Upon initial setup, the System Admin creates:

  • IA Domains. A Domain is based on a single chatbot interface. A Domain usually serves a unique area of user interest and is created according to the needs of your service goals and your site. Example Domains might include "Travel" or "Billing Support". You can have multiple Domains in your Account, but each Interaction entry point can have only one Domain.
  • User Languages. Each IA Domain can use multiple languages concurrently. Each language is maintained using a separate set of NLP components.

After setup, the Account Admin creates the basic design components of the IA, including:

  • NLP Components. The Intelligent Assistant uses advanced NLP to determine precisely what a user wants to do at any point of a workflow. NLP capabilities operate on the basis of:
    • Intents. What the user is looking for or trying to do. An Intent represents a mapping between what the user requests and the resulting navigation in Interact. Each Intent is associated with an approved Interaction which is run when the NLP Engine recognizes the Intent from the user's request.
    • Entities. A set of pre-defined characteristics that help to clarify the user's Intent. An Entity is usually mapped to a user selection or a specific variable in an Interaction.
    For a detailed explanation, see About Intents and Entities.
  • Training Sentences. The NLP engine is trained by providing initial training sentences to each new Intent and mapping the Entities that are used. As real time user requests are continuously collected for a Domain, NLP engine performance is improved by frequently reviewing and correcting engine assignments. The more training that is performed for a Domain engine, the more knowledge will have about possible user requests and needs.
  • Testing. Especially during design activities, potential user request sentences can be tested to see how the NLP engine responds. Testing returns the Intent that the NLP engine would assign, the probability for its successful assignment, and also the Entity parameters that were parsed from the sentence.

Using Multiple Languages

The IA and its components support use of more than one language concurrently. When setting up an IA Domain for your account, the System Admin will also set up the additional languages you have requested.

  • For each Intent, the Account Admin can create training sentences in multiple languages. For each Entity, the Account Admin can create values in multiple languages.
  • Testing sentences on the NLP engine can also be performed in any of the languages that have been set up for the Domain.
  • When reviewing user requests during Training, the Account Admin assigns the appropriate Language, along with the Intent for each.

Adding Conversational Chatbots

When the basic design of the IA Domain is completed, the Designer creates IA text entry chatbots – just like any other page element – on the desired pages of their Interactions. During creation, each chatbot is associated with a specific Domain. For details, see Adding the Intelligent Assistant to an Interaction.

During runtime the NLP engine analyzes user requests from the chatbot to determine the Intent, and then uses the recognized Intent's parameters to help provide solutions via existing workflows.

Resources Created Automatically by the IA

The IA transparently creates Custom Type variables to store key IA components for each Intent, such as user input and NLP responses. A Designer can then insert the Custom Type as a local variable for use in an Interaction.

You can view IA-related Custom Types as read-only values anytime in the Admin Console.