Open conversational ai framework for building contextual chatbots and ai assistants

Copyright © 2019 HCL Technologies Limited | www.hcltech.com
OPEN CONVERSATIONAL AI FRAMEWORK FOR BUILDING CONTEXTUAL CHATBOTS AND
AI ASSISTANTS.
BY - PRASANNA VENKATESH JAYAPRAKASH (PRASANNAVJ@HCL.COM)

2 Copyright © 2019 HCL Technologies Limited | www.hcltech.com
AGENDA
Conversational AI Framework for building contextual chatbots and AI assistants.
 Understanding Chatbot Architecture
 Introduction to Current Chatbot Platforms and Industry Problem
 Introduction to RASA Framework
 introduction to RASA Chatbot Architecture
 Natural Language understanding - Components
 CORE Components and Structure.
 Core - Stories and Dialog management
 Integrating with External API
 Demonstration using Restaurant Details capture details.
 Demonstration integrating with Slack Messaging Platform.

CHATBOT ARCHITECTURE

INTRODUCTION TO CURRENT CHATBOT PLATFORMS AND INDUSTRY PROBLEM
 Google dialogFlow
 IBM
 Microsoft
 Amazon
 Facebook
 Industry problem:
 To Implement the solution no inhouse platform.
 All depend on cloud based learning solution – when enterprise plan to integrate might face problem due to the security and privacy concern of the
providers.
 Old rule based providers are phasing out
 Solution:
 Open Source conversational platform enable building and managing the complete chatbot inhouse – one such platform is RASA.

INTRODUCTION TO RASA FRAMEWORK
 Utilizes customizable ML and NLP solution to enable developers to control how and what they wanted to achieve part of the conversation.
 Provide option to integrate with external API services, External databases etc.
 Provide option to integrate with various prominent chatting/Messaging solution i.e: - Slack, faceBook messenger etc.
 Provide both Open Source (RASA Stack) and Commercial Offering (RASA Platform)

INTRODUCTION TO RASA CHATBOT ARCHITECTURE

NATURAL LANGUAGE UNDERSTANDING - COMPONENTS
Creating using
 Entity
 Intents
 Using RASA NLU Trainer

RASA CORE COMPONENTS AND STRUCTURE - DIALOG MANAGEMENT
 Entity
 Intents
 Slots
 Templates
 Actions
 Buttons

ENTITY
- Entity allows us to
define the list of items that
Need to be recognized by
Chatbot while conversing

INTENTS
- Intent of the person
Interacting with the chatbot

SLOTS
- Recognized by chatbot and
filled by the bot for further
processing.

ACTIONS
- Action to be carried out by
Chatbot on specific input
- Searching for restaurant
- Connecting with back end
Db etc.

TEMPLATES
- Possible options for display
To the used based on the
Specific intent

DIALOG MANAGEMENT MODULE - STORIES
 Stories
- Potential option for
conversation flow is detailed
In the conversation.

CONVERSATION EXAMPLE - DEMO
 Demo of solution:
 Video Link: https://www.youtube.com/watch?v=qi0o3jwMjvo&t=26s
 Live demo will be conducted during the presentation session.
 Starting from setting up the NLU
 Setting up core
 Integrating with Zomato API
 Integrating with SLACK

THANK YOU
PRASANNA VENKATESH JAYAPRAKASH
PRASANNAVJ@HCL.COM
LINKEDIN – HTTPS://WWW.LINKEDIN.COM/IN/PRASANNA-VENKATESH-JAYAPRAKASH-454133A/
TWITTER - @PRASANNA_VRI

Open conversational ai framework for building contextual chatbots and ai assistants

More Related Content

Similar to Open conversational ai framework for building contextual chatbots and ai assistants (20)

Recently uploaded (20)

Open conversational ai framework for building contextual chatbots and ai assistants