Design Conversational AI Experience

In this section, we will introduce you to a good conversation with an agent. In the process, we will explain to you what is an idea conversation and what can you expect from a well-designed conversational AI solution.

In this section, we will cover key Conversational AI concepts, specifically intents and utterances.   We will cover definition of Intents and Utterances and will use several examples to explain these concepts

In the lecture we will cover key conversational concept - Entities.  We will provide its definition and several  examples.

Conversations are often done in a context that must be tracked during a conversation session or across multiple conversation sessions between two parties. In this section we will cover Context with several examples

Dialog in a conversation determines how sentences are sequenced.  In this lecture, we will cover Dailog management with examples.

In this lecture, we will cover short tail and long tail questions

We are naturally guarded in our conversations whether dealing with humans or machines. This is especially true if a stranger were to start reciting confidential information and aggressively sell products. On the other hand we easily confide when our chemistry clicks.

In this section, we will cover different user personas and how conversational AI system should be designed to reflect and adjust to user personas

In this section we will cover different classes of Conversations AI engines like machine to machine or Robotic Assistant or Virtual Agent. We will cover machine to machine chatbot in this lecture

In this lecture we will cover Robotic Assistant. A Robotic Assistant is an apprentice role for providing system queries and performing searches and computations, but not lead the conversation.

In this lecture we will cover Virtual Agent.  In a virtual agent experience, Robot is responsible for carrying out conversation to help with any specific expertise area.

In this section we will cover how you measure complexity for a Conversational AI system. We will take six major dimensions. For each of these six dimensions, we will explore levels of sophistication and provide you with an appreciation for how different conversational AI agents can be designed with varying levels of sophistication. We will apply these dimensions to create Conversational AI Agents at 4 levels of sophistication. Most use cases can be addressed via either of these four levels. 

In this particular lecture, we will cover Level 1- Structured Query and Response.

In this lecture, we will cover Level 2 complexity level - Natural language query and response

In this lecture, we will cover Level 3 is Reactive Context-sensitive conversation bot.. It uses unstructured query and unstructured response with disambiguation. It has Ability to use context and disambiguation to listen and interpret and non-verbal cues in proactive or reactive situation

In this lecture, we will cover Level 4  goal-directed consultative conversation. It includes goal-directed proactive conversation with response variation to match robot personality, customer empathy management and non-verbal cues

Now that we have introduced the Conversational AI system, it is time to build. In this section, we will introduce various solution components of a Conversational AI system and how they are stacked together for an overall solution.

In this lecture, we will provide overview of each of the 5 sub-systems associated with 1) End User Interfaces, 2) Training System, 3) Run time system, 4) Library Management, 5) Analytics and Reporting sub-systems.

In this lecture, we will first describe a use case.  Then we will design key concepts of a Conversational AI solution such as Utterances, Intents, Entities on this use case example

In this lecture, we will design several conversational flow design with different system chemistry

This lecture uses few slides to summarize key contents and remind you of some of the key learnings in this class so that you can test drive it.

This lecture summarizes future courses and learnings in the field of AI from Applied AI Institute