Build Advanced Enterprise Chatbots, Voice, and Telephony Agents on Google Cloud

Table of Contents
About the Author
About the Technical Reviewers
Acknowledgments
Introduction
Chapter 1: Introduction to Conversational AI
	The History of Text Chatbots
	Why Do Some Chatbots Fail?
	Machine Learning Simply Explained
		Natural Language Processing
	Chatbots and Artificial Intelligence
	Machine Learning and Google
	About Google Cloud
		Open Source
	About Dialogflow
		Dialogflow Essentials and Dialogflow CX
			How Dialogflow Essentials Works
			How the Industry Is Changing Its Complexity
			Where Dialogflow CX Fits In
			Dialogflow CX Explained
			Dialogflow Essentials vs. Dialogflow CX
	About Contact Center AI
		CCAI Architecture
	About Google Cloud Speech Technology
		Cloud Speech-to-Text API
		Cloud Text-to-Speech API
		WaveNet
		Custom Voice
	Other Google Conversational AI Products
		Google Assistant
		Actions on Google
			Actions Builder
		AdLingo
		Chatbase
		Duplex
		Meena & LaMDA
	Summary
	Further Reading
Chapter 2: Getting Started with Dialogflow Essentials
	Dialogflow Essentials Editions
	Creating a Dialogflow Trial Agent
	Creating Dialogflow Agents for Enterprises
		Quotas
		User Roles and Monitoring
			Using VPC Service Controls
		Using Developer Features
	Configuring Your Dialogflow Project
		General
		Languages
		ML Settings
			Automatic Spell Correction
			Automatic Training
			Agent Validation
		Export and Import
		Environments
		Speech
			Improve Speech Recognition Quality
				Enable Enhanced Speech Models and Data Logging
				Enable Auto Speech Adaptation
			Text to Speech
				Enable Automatic Text to Speech
				Voice Configuration
		Share
		Advanced
	Configuring Your Dialogflow for Developers
	Summary
	Further Reading
Chapter 3: Dialogflow Essentials Concepts
	Intents in Depth
		Setting Up Intents
	Entities in Depth
		Creating Custom Entities
		Advanced Custom Entities
	Creating Intents with Entities in Training Phrases
	Keeping Context
		Setting Up Follow-Up Intents
		Manually Setting Input and Output in “Normal” Intents
		Lifespan
		Keeping Context with the SDK
	Testing in the Simulator
	Summary
	Further Reading
Chapter 4: Building Chatbots with Templates
	Creating Prebuilt Agents
	Enabling Small Talk Modules
	Creating an FAQ Knowledge Base
	Best Practices
		Convert Knowledge Base Questions to Intents
	Summary
	Further Reading
Chapter 5: Bot Management
	Agent Validation
		Understanding Validation Results
		Validation via the SDK
	Improve the Dialogflow Machine Learning Model with Built-in Training
	Summary
	Further Reading
Chapter 6: Deploying Your Chatbot to Web and Social Media Channels
	Integrating Your Agent with Google Chat
		Enabling Your Agent in the Google Chat
		Rich Messages Support
		More Text-Based/Open Source Integration Options
	Integrating Your Agent with a Web Demo
	Integrating Your Agent with a Dialogflow Messenger
		Changing the Look and Feel of the Chatbot Component
		Rich Messages Support
	Summary
	Further Reading
Chapter 7: Building Voice Agents
	Building a Voice AI for a Virtual Assistant Like the Google Assistant
		Rich Messages
		Fulfillment and Webhooks
		Invoke Your Action on the Google Assistant with Explicit and Implicit Invocation
		Submitting an Action via Actions on Google
		Building an Action with the Actions SDK
			Using the Actions SDK Solution
			Deploying Your Action
	Building a Callbot with a Phone Gateway
		Response Messages for the Phone Gateway
	Building Bots for Contact Centers with Contact Center AI
		Enabling Contact Center AI
	Improving Speech to Text Quality
		Custom Entities Hints
		System Entities Hints
		Intent Hints
		Overriding Speech Hints in Your Code
	Fine-Tuning the Text to Speech Output of Voice Bots with SSML
		UX Design for Voice Dialogues Matters!
		Text to Speech Voices
		Controlling the Intonation
	Summary
	Further Reading
Chapter 8: Creating a Multilingual Chatbot
	Agent Languages
	Building a Multi-language Agent
	Exporting a Multi-language Dialogflow ES Agent
	Detecting Multi-language Intents via the SDK
	Working with the Translation Service
	Summary
	Further Reading
Chapter 9: Orchestrate Multiple Sub-chatbots from One Chat Interface
	Creating a Mega-Agent
	Using the SDK
	How Billing Works
	Summary
	Further Reading
Chapter 10: Creating Fulfillment Webhooks
	An Introduction to Fulfillment Webhooks
	Building a Fulfillment with the Built-in Editor
		Enable Fulfillment
		Using the dialogflow-fulfillment Package
			Diagnostic Info
			Firebase Logs
		Using Actions on Google for Building Dialogflow Fulfillment
		Build Your Fulfillment Webhook Manually
	Building Fulfillments Webhook
		Where to Run My Back-End Code?
			Cloud Functions
			App Engine (Flexible Environment)
			Cloud Run
			Kubernetes Engine
			Compute Engine
		Enable Webhooks
		Cloud Function Implementation
		Express Implementation (with Cloud Run)
		Google Cloud Logging
	Building Multilingual Fulfillment Webhook
		i18n Code Example
	Using Local Webhooks
		Ngrok
		Testing Your Fulfillment Without Dialogflow and ngrok
	Securing Webhooks
		Basic Authentication
		Authentication with Authentication Headers
		Mutual TLS Authentication
			Valid Secure SSL Certificate
			Root CA
			HTTPS Authentication Setup with Apache
			A Full Example for Setting Up Mutual TLS Authentication
				Create a Node.js VM on Compute Engine
				Attach a Domain Name to Your VM
				Set Up Your Node Application
				Set Up mTLS
	Summary
	Further Reading
Chapter 11: Creating a Custom Integration with the Dialogflow SDK
	Implementing a Custom Chatbot in Your Website Front End, Setup
	UI Implementation
	Back-End Implementation
	Welcome Message
	Creating Rich Responses in Your Chatbot Integration
		A Hyperlink Component, a Google Map, and an Image Component
			Implementation
	Using Markdown Syntax and Conditional Templates in Your Dialogflow Responses
		Branching the Conversation
	Building an Integration to Run a Dialogflow Agent in a Native Mobile Android or iOS App with Flutter
		Two Techniques for Integrating Dialogflow in a Flutter Application
		Integrating the Dialogflow SDK Directly into Your Flutter App
		A Flutter App That Communicates with a Back-End Dialogflow SDK App
	Summary
	Further Reading
Chapter 12: Implementing a Dialogflow Voice Agent in Your Website or App Using the SDK
	Reasons for Not Picking Google Assistant
	Building a Client-Side Web Application That Streams Audio from a Browser Microphone to a Server
		Build the Front End
		Short Utterance vs. Streaming
		Record Single Utterances
		Record Audio Streams
	Building a Web Server That Receives a Browser Microphone Stream to Detect Intents
		Dialogflow vs. Text-to-Speech API vs. Speech-to-Text API
			Speech-to-Text API
			Text-to-Speech API
		Build the Back End
			API Calls to Dialogflow
			DetectIntent
			StreamingDetectIntent
	Retrieving Audio Results from Dialogflow and Playing It in Your Browser
		Client-Side Code to Play the Audio
	Summary
	Further Reading
Chapter 13: Collecting and Monitoring Conversational Analytics
	Conversation-Related Metrics
	Customer Rating Metrics
	Chat Session and Funnel Metrics
	Bot Model Health Metrics
	Capturing Conversation-Related Metrics to Store in BigQuery
		BigQuery
		Capture Points
			Session Id
			Date/Timestamp
			Sentiment Score
			Language and Keyword
			Platform
			Intent Detection
		Solutions
			Building a Platform for Capturing Conversation-Related Metrics and Redacting Sensitive Information
			Detecting User Sentiment
			Topic Mining
	Collecting Customer Rating Metrics
		Net Promoter Score (NPS)
		Customer Satisfaction (CSAT)
		Customer Effort Score (CES)
	Monitoring Chat Session and Funnel Metrics with Dialogflow, Chatbase, or Actions on Google
		Metrics to Monitor
			Total Usage
			Percentage of Users That Matches the Intent
			Completion Rate
			Drop-Off Rate/Drop-Off Place
		Channel-Specific Metrics to Monitor
			User Retention
			Endpoint Health
			Discovery
		Dialogflow Built-in Analytics
		Monitoring Metrics with Chatbase
		Analytics on Actions on Google
	Capturing Chatbot Model Health Metrics for Testing the Underlying NLU Model Quality
		True Positive—A Correctly Matched Intent
		True Negative—An Unsupported Request
		False Positive—A Misunderstood Request
		False Negative—A Missed Request
		True Positive Rate
		False Positive Rate
		ROC Curve
		Accuracy
		Precision
		F1 Score
		Confusion Matrix
	Summary
	Further Reading
Appendix: An Introduction to Dialogflow CX
	How the Industry Is Changing Its Conversation Complexity
	Where Dialogflow CX Fits In
	Dialogflow CX Features
	New Concepts in Dialogflow CX
		Flows
		Pages
		State Handlers
		Contact Center Features
		Customer-Managed Encryption Keys (CMEKs)
	Where Both Products Differ
		Agents
		NLU
		Analytics
		Entities
		Intents
		Fulfillment and Webhooks
		APIs
		Error Handling
	When to Use Dialogflow CX vs. Dialogflow ES?
	Summary
	Further Reading
Index