 |
|
|
|
Dynamic
responses
|
The Existor conversational AI chooses its responses dynamically based on user input. It is dynamic in several ways.
Often scripting will aim to lead a conversation, such as by asking appropriate questions and predicting the answers it will receive. The AI makes a selective judgement as to the most appropriate response based on the context, including everything said by and known about the user.
The AI employs fuzzy logic, yet with controllable accuracy requirements for each case. People talking with our AI Avatar may themselves lead, interrupting its 'script' whenever they like with asides, requests for explanations and more, and the AI will get back on track after handling them.
The wording of every sentence of the AI's responses is varied as a reflection of a person's style and the information they've given - potentially to such a degree that no two users would ever see the same thing.
On a larger scale, the AI learns from user input and re-evaluates, according to accumulated experience, the value of its future actions relative to the goals and objectives defined for it. Conversations that lead to successful outcomes will be favoured in similar circumstances, a positive feedback-loop of self-improvement.
This dynamism, along with genuine understanding, sets our AI apart. Dynamic responsivness and clear displays of comprehension and memory make it seem more human, believable, engaging and even entertaining.
Allied with superior, contextual visual and audio gestures from avatars, the promise of AI can now be delivered.
|
|
|
