Build your own chatbot in just 25 minutes!

Table of Contents:

1. Introduction
2. Accessing the Bot Builder
3. Creating Bots on the Core Platform
   1. Creating a Standard Bot
   2. Creating a Soldad Bot
   3. Bot Configuration and Tasks a. Dialogue Tasks b. Knowledge Tasks c. Alert Tasks d. Information Tasks e. Action Tasks
4. Designing the Conversation Flow
   1. Dialogue Builder
   2. Node Properties
   3. Entities and Services
   4. Script Logic and Customization
   5. Confirmation and Messaging
5. Testing and Training the Bot
   1. Natural Language Engine
   2. Machine Learning Model
   3. Synonyms and Patterns
   4. Fallback Dialogue
   5. Utterance Testing and Batch Testing
6. Enabling and Publishing the Bot
   1. Channel Configuration
   2. Bot Kit SDK and Custom Logic
   3. Publishing Changes and Approval
7. Analyzing and Optimizing Performance
   1. Intent Recognition and Training Data
   2. Failed Utterances and Aborted Tasks
   3. Performance Metrics and Logs
8. Bot Configuration Settings
   1. Name, Logo, and Welcome Message
   2. Language Support and Collaboration
   3. Importing and Exporting Configuration
   4. IVR and Advanced Channel Settings
   5. Change Logs and Dashboard Metrics

Article:

Introduction

Welcome to the world of bot building on the Core platform! This comprehensive guide will take You through the process of creating, designing, and publishing your own bots. Whether you're a beginner or an experienced developer, this article will provide you with all the information you need to get started and optimize the performance of your bots.

Accessing the Bot Builder

To access the Bot Builder, you need to go to the Developer section of the Core platform. Once you're there, click on the Bot Builder icon located in the top right corner. The Bot Builder is a web-Based IDE that allows you to design, develop, and deploy bots on the Core platform.

Creating Bots on the Core Platform

When you enter the Bot Builder, you'll see a list of bots that you've developed or that have been shared with you by other developers. You can start by creating a brand new bot called a Standard Bot or build a more complex bot called a Soldad Bot, which utilizes multiple bots as part of a single board configuration.

Bot Configuration and Tasks

In the bot configuration, you'll Notice that your first task is the Dialogue Task. This task represents the functional behavior of the bot and is where you'll define the conversation flow. Additionally, you'll have other tasks such as Knowledge Tasks, Alert Tasks, Information Tasks, and Action Tasks.

Dialogue Tasks

The Dialogue Task is the heart of your bot's conversation flow. In the Dialogue Builder, you'll have a visual drag-and-drop environment to describe the steps and interactions between the user and the bot. Each node in the dialogue represents a specific action or question, and you can set various properties for each node.

Knowledge Tasks

The Knowledge Task is used when you want the bot to be able to answer frequently asked questions. You can import questions and train the bot to provide accurate answers without the need for backend lookups. The Knowledge Task allows you to organize the questions based on an ontology structure and configure alternate phrases for each question.

Alert Tasks

The Alert Task is useful when you want the bot to proactively engage the user in a conversation based on user-defined events or enterprise-defined alerts. You can configure various parameters for the alert and define the actions the bot should take when triggered.

Information Tasks

The Information Task allows the user to ask or Seek information from backend systems using natural language. For example, a user can ask for transaction details or search for specific information. The platform converts the natural language input into an information query and presents the response to the user.

Action Tasks

The Action Task is a slot filling task where the bot may ask specific questions or slots to fill before performing a certain action. For example, reporting a lost or stolen card might require the bot to ask for details such as the card number and reason. Once the slots are filled, the bot can perform the requested action.

Designing the Conversation Flow

In the Dialogue Builder, you can define the conversation flow of your bot by connecting the nodes and specifying the conditions for transitioning between nodes. Each node has properties that you can set, such as intents, entities, and services.

Node Properties

Each node in the Dialogue Builder has various properties that you can configure. For example, the Intent Node has properties related to the intent it represents, while the Entity From Node has properties related to the entity it represents. You can also add service nodes to make API calls or provide script logic for custom business rules.

Entities and Services

Entities play a crucial role in recognizing different forms of user inputs. The platform provides a set of pre-configured entities that you can choose from, or you can Create custom entities based on the specific needs of your business. Service nodes allow you to make API calls and configure the necessary parameters and headers.

Script Logic and Customization

You can write script logic in the Dialogue Builder to perform various business operations, such as preparing a payee list or validating data before making a payment. The platform supports JavaScript logic, allowing you to dynamically generate messages and customize the conversation based on the data available.

Confirmation and Messaging

Confirmation nodes allow you to validate user input before proceeding with a transaction. You can configure the conditions for confirming or rejecting the transaction and perform additional actions based on the user's response. Messages can be customized using plain text or JavaScript logic and tailored for different channels or user interaction scenarios.

Testing and Training the Bot

Once you have defined the conversation flow and configured the tasks, it's time to test and train your bot for accurate intent recognition and response generation.

Natural Language Engine

The Core platform offers multiple natural language engines for intent recognition. You can provide training data to each engine to improve its accuracy in identifying intents and entities. The platform allows you to Visualize the scoring and performance of your training data, helping you identify any issues and optimize your model.

Machine Learning Model

The machine learning model in the Core platform uses deep neural networks to score and map user utterances to intents. You can provide utterances for various intents to train the model. The platform also provides sentiment analysis for user messages, helping you gauge user sentiment throughout the conversation.

Synonyms and Patterns

The natural language engine uses synonyms and patterns to recognize intents and entities accurately. While the platform provides a set of pre-configured synonyms and patterns, you can add custom ones to handle unique terminology or user expressions. This customization enhances the engine's ability to understand user inputs.

Fallback Dialogue

In case the bot fails to recognize an intent, a fallback dialogue can be configured to handle such scenarios. The fallback dialogue acts as a default response for unrecognized user inputs, ensuring a smooth conversation flow even in unexpected situations.

Utterance Testing and Batch Testing

You can test individual utterances and batches of utterances to evaluate how well your bot recognizes intents and generates responses in different scenarios. The platform allows you to analyze the scoring, match results, and sentiment of each utterance. Batch testing helps you validate changes to the natural language model without affecting previously successful user interactions.

Enabling and Publishing the Bot

Once you're satisfied with the design and performance of your bot, it's time to enable it in various channels and publish the changes to make the bot available to users.

Channel Configuration

The Core platform supports multiple channels, such as web, mobile, SMS, and social media. You can enable specific channels by following the configuration steps provided by the platform. The channel configuration allows you to optimize the messages and responses based on the capabilities of each channel.

Bot Kit SDK and Custom Logic

For more granular control and customization, the Core platform provides Bot Kit SDKs. These SDKs allow you to introduce custom logic and execution paths to meet the unique requirements of your business. The Apps and Agents view is where you can configure the Bot Kit SDK and other capabilities of the platform.

Publishing Changes and Approval

To make your bot changes live, you need to publish the configuration. An administrator can review and approve the changes before they are pushed into the production environment. This approval process ensures that only tested and approved changes are made available to users.

Analyzing and Optimizing Performance

Once your bot is live, it's essential to monitor and analyze its performance to ensure a smooth user experience. The Core platform provides various tools and metrics to help you optimize your bot's performance.

Intent Recognition and Training Data

The platform tracks the performance of intent recognition for each user utterance. You can review and train the bot if an intent should have been matched but wasn't. By analyzing utterances and training data, you can continuously improve the accuracy and responsiveness of your bot.

Failed Utterances and Aborted Tasks

The Failed Test App allows you to analyze why a user aborted a task or why a particular transaction failed. By studying the chat history and Context, you can identify the reasons for user frustration or unsuccessful task completion. This information can be used to optimize the dialogue and enhance the overall bot experience.

Performance Metrics and Logs

The Dashboard view provides a comprehensive overview of your bot's performance metrics. You can see the number of tasks performed over time, peak usage periods, intent accuracy, and popularity of different tasks and channels. Additionally, you can review the change logs to keep track of all the modifications made to your bot configuration.

Bot Configuration Settings

The Settings tab allows you to configure various aspects of your bot, including its name, logo, welcome message, and supported languages. You can invite other developers to collaborate on the bot, define bot variables for environment transitions, import/export configurations, and configure advanced settings such as IVR and channel-specific preferences.

Conclusion

Building and optimizing bots on the Core platform is an exciting endeavor. By following the step-by-step guide provided in this article, you can create highly functional and user-friendly bots that deliver exceptional experiences to your customers. Remember to continuously monitor your bot's performance, analyze user interactions, and make necessary improvements to ensure an efficient and engaging bot experience.

Highlights:

• The Bot Builder on the Core platform allows developers to design, develop, and deploy bots.
• The Bot Builder offers a visual drag-and-drop environment for designing the conversation flow.
• Various task types, such as dialogue, knowledge, alert, information, and action tasks, can be used to define bot behavior.
• The Dialogue Builder in the Bot Builder enables developers to create and customize the conversation flow.
• The Natural Language Engine and Machine Learning Model aid in accurate intent recognition and response generation.
• Synonyms and patterns can be added to improve the recognition of user inputs.
• Fallback dialogues ensure a smooth conversation flow in case of unrecognized intents.
• Batch testing allows developers to validate changes to the natural language model.
• The Bot Builder supports multiple channels and provides customization options for each channel.
• Performance metrics and logs help in analyzing and optimizing the bot's performance.
• The Settings tab in the Bot Builder allows developers to configure various aspects of the bot.

FAQ:

Q: What is the Bot Builder on the Core platform? A: The Bot Builder is a web-based IDE that allows developers to design, develop, and deploy bots on the Core platform.

Q: How can I design the conversation flow of my bot? A: The Dialogue Builder in the Bot Builder provides a visual drag-and-drop environment for designing the conversation flow. You can connect nodes and define conditions for transitioning between them.

Q: What are the different task types available in the Bot Builder? A: The Bot Builder supports various task types, including dialogue tasks, knowledge tasks, alert tasks, information tasks, and action tasks.

Q: How can I improve the accuracy of intent recognition in my bot? A: By providing training data to the natural language engine and machine learning model, you can improve the accuracy of intent recognition. Additionally, adding synonyms and patterns specific to your business can enhance recognition.

Q: Can I customize the responses based on the channel? A: Yes, the Bot Builder allows you to customize messages and responses based on the capabilities of each channel.

Q: How can I monitor and analyze the performance of my bot? A: The Core platform provides performance metrics, logs, and tools to analyze the bot's performance. You can track intent recognition, review failed utterances, and monitor various performance metrics on the dashboard.

Q: Can I collaborate with other developers on bot development? A: Yes, the Bot Builder allows you to invite other developers to collaborate on bot development. You can work together and make changes to the bot configuration.

Q: What configuration options are available for the bot? A: The Settings tab in the Bot Builder allows you to configure various aspects of the bot, including its name, logo, welcome message, supported languages, collaboration options, and advanced settings.

Q: How can I publish changes made to the bot configuration? A: To publish changes, you can submit them for review and approval by an administrator. Once approved, the changes will be pushed into the production environment.

Q: How can I import and export bot configuration? A: The Core platform allows you to import and export bot configuration using CSV or JSON files. This allows you to efficiently transfer configurations between different environments or repositories.

Q: What are the benefits of using the Bot Builder on the Core platform? A: The Bot Builder on the Core platform provides a comprehensive set of tools and features for designing, developing, testing, and optimizing bots. It offers a visual interface, customizable conversation flow, training and testing capabilities, multiple channel support, performance analysis tools, and advanced configuration options.