Discourse Engine™
The Discourse Engine is a rules-based artificial intelligence engine for integrating intelligent agent personalities into training, performance support, and entertainment applications. This market-leading tool is suitable for a wide variety of applications.
It supports conversation mode for simulating conversations with one or more intelligent agents, master-apprentice mode for reviewing deliverables and providing intelligent guidance, and auto mode for mass-customization of correspondence documents in production applications.
Current Features
- Intelligent Agent Personalities - Create a personality, teach it about the content area, the reference materials, and the learner's goals, make it react to the learner accordingly. Very fine control over interaction. Can simulate 360° feedback (supervisor, peer, subordinate, adversary). Can have many personalities in Sim.
- Object-Only Authoring (No Code) - Intelligent feedback/conversation capability can be created and tested by non-technical team members without writing code or assistance from coders.
- COM object/ActiveX Control - The Discourse Engine is a COM object that can be used and called by any COM-compliant development tool(VB, VC, VBA, IE4+ web pages). It can also be wrapped by an ActiveX control or Add-in for easy integration into different environments. (For example, SheetSmarts is a COM Add-In for connecting the Discourse Engine with Excel).
- Hooks to wrapping applications (Black Boxes and Sensors) - The algorithm in the Discourse Engine contains objects to define hooks to pass data to and from the calling application. We have a 'BlackBox' that models a calculations engine such as Powersim, Excel, or a custom-coded module. It also has a 'Sensor' object that hooks to individual data items.
- Exposed Object API - The API exposes not just functions for calling built-in functionality, but also the objects and collections themselves, so additional functionality can be added externally, if desired.
- Rules based analysis - We have an intuitive, generic, rule-based way to analyze the learner's decisions.
- Nestable Rule Objects - A Rule's logic can be reused by reference in other Rules. (ie RuleC should fire if RuleA and RuleB fired)
- Flexible Feedback Object Assembly Algorithm - Feedback is objects, not just text. Personalities can be configured to give a lot or a little feedback. Multiple pieces of feedback can be assembled based on the user's specific needs. Feedback can consist of text, html, xml, follow up questions, events that trigger actions in the application, and links to reference materials.
- Text Substitution Language - Feedback can dynamically determine the problem the user is having and give the right answer, or not.
- Custom Extensibility/ Application Event Model - Feedback can contain generic Event objects that applications can handle in a custom way.
- User Deliverable Modeling - The algorithm includes objects that model the decisions that the user will make.
- Java Integration - The Discourse Engine has a Java wrapper. This allows us to use the get the benefit of C++ speed in Java based environments.