COOPERSTOW

N DREAMS
PARK
CHAT BOT METHODOLOGY
 Contents

1. Google’s Dialogue flow 2. Amazon’s LEX 3. Microsoft’s Bot Frame work

A. Work Flow A. Work Flow A. Work Flow
B. Advantages B. Advantages B. Advantages
C. Disadvantages C. Disadvantages C. Disadvantages
 Dialogflow

01 Types Methodology 02

03 Advantages Disadvantages 04
 Dialogflow offers two main types of agents:
Dialogflow E S (Essentials) and Dialogflow C X
(Customer Experience).

Types ES is used for simple chatbots like FAQ bots, small

business assistants, or educational projects.

CX is used for enterprise solutions where

conversations are complex and dynamic, such as
contacts. centers or large-scale customer support
system
 1. Define Objectives
2. Prepare project setup
3. Create a Dialogflow Agent

Methodolog 4. Define Intents

5. Leverage Entities
y 6. Implement Contexts
7. Configure Fulfillment
8. Test the Chatbot
9. Integrate the Chatbot
10.Maintain and Scale
 Identify the purpose of the
chatbot, such as customer support or FAQs.

Define State who the audience will be and the type

of questions to ask.

Objectives Set measurable goals, such

as decreasing customer response time
or increasing user engagement.
 Make a Google cloud platform account.
Prepare Activate Dialogflow API in project using google
project cloud console

setup
 Agent Name
Create a Default Language
Time zone
Dialogflow Enviournment

Agent
 Intents are actions which our chatbot will perform
based on user’s query.

Example : Registration, Food facility

Intents
Steps:

Create Intents
Provide Examples
Provide Response
Set Parameters and entities
 2 type of entities are available.

Entities System entities like date and time.

Custom entities like team name.

 Enabling webhook fulfilment is to connect
dialogflow with our backend system for
dynamic responses.
Fulfillment
Prepare a framework using flask or node.js
Process requests from dialogflow and return
custom replies

Real time data fetching

Trigger external action
 Check the following thing:

Is each intent is fulfilled

Test Is user friendly

chatbot
 Select the platform on which we would like to
deploy our chatbot.

Integrate
chatbot
 Update chatbot with latest intent and entities
as the chatbot evolves.

Maintain
and Scale
Advantage
s1. Cost Effective : Suitable for small and medium-scale businesses with
affordable pricing tiers.

2. Advanced NLP Capabilities: Uses Google’s powerful Natural Language

Processing (NLP) to understand user inputs with high accuracy. Recognizes
intents and entities effectively, improving user interactions.

3. Built-In Analytics: Provides insights into user integrations, intent usage and
performance metrics.
Disadvantage
s1. Scalability Challenges
2. Limited Role Management
3. No version control
4. Lack of collaboration
 Amazon’s Lex

01 Methodology

Advantages 02
03 Disadvantages
 1. Define the bot’s Purpose and Scope
2. Design the conversational Flow

Methodology 3. Build the bot using amazon lex

console
4. Train the bot
5. Integrate with external services
6. Test the bot
7. Deploy the bot
8. Monitor and optimize
9. Update as needed
Define the bot’s purpose and
scope

1. Goal Setting: Identify the objective of chatbot

2. Target Audience: Text based, voice based or both.
Design the conversation flow
Intent Definition : Define the intents that represent the tasks that bot should
handle.

Entities Definition: Identify entities that represent the specific information

needed for each intent.

Dialog flow: Design a conversational flow for each intent , defining the questions
the bot will ask and the responses it will give

Context Management: Plan how the bot will manage conversation context.
Amazon Lex
Create a new bot: Give bot name, language , text based or voice based or both if
voice based provide output voice.

Define intents: For each use case create an intent with sample user phrases and
define expected responses.

Set up slots: Set up slots that the bot will collect from users.

Configure prompts and validation: Set prompts for the bot to ask the user for
missing information

Fulfillment setup:Set up AWS Lambda functions for backend logic or data

retrieval
Train the bot
Training Phrases: Add sample user inputs (training phrases) that represent how
users might ask questions or issue commands related to each intent.

Test the Bot: Use the built-in testing tools in Amazon Lex to simulate
conversations and test the bot’s response to user inputs.

Iteration: Refine the bot’s responses, intent recognition, and dialog flow based on
testing results and feedback.
Integrate with External Services

AWS Lambda Integration:Integrate with AWS Lambda for any dynamic backend
processing

External API Integration: Use Lex’s built-in webhooks or Lambda functions to

interact with external APIs or databases for real-time data fetching.
Test the Bot

Simulate Conversations: Use the Lex Console’s Test Chat feature to simulate
interactions, test intents, and validate that the bot provides correct responses.

Real-World Testing: Conduct user testing by having real users interact with the bot
to identify edge cases and areas for improvement.

Context Handling: Test how well the bot maintains conversation state and context
over multiple interactions.
Deploy the bot

Amazon Lex can be integrated into various messaging platforms.

Deploy the bot to a custom web page or mobile app using SDKs for text based
interaction.
Monitor and Optimize

Analytics: Use Lex Analytics to track user interactions, identify common

questions, and determine any failure points (e.g., intents with poor recognition).

Performance Optimization: Based on usage data, adjust intents, improve

training phrases, refine response patterns, and optimize Lambda function
performance.

Continuous Improvement: Collect feedback from users to further improve

conversational design and AI capabilities over time.
Update the Bot as Needed

New Features: Continuously add new intents, update training data, or modify
workflows as the bot’s functionality grows.

Re-Training: Re-train the bot periodically to improve intent recognition and

make it more robust against user input variations.

Version Control: Use versioning within the Lex Console to manage different
versions of the bot and deploy updates safely.
 Advantages

1.Seamless Integration with AWS Services: Amazon Lex integrates easily with
AWS services like Lambda so we can directly retrieve data from our database

2.Natural Language Processing and Speech Recognition: Lex utilizes advanced

NLP and ASR to process natural language and voice input.

3.Cost-Effective and Scalable: With a pay-as-you-go model and automatic scaling,

Lex offers cost efficiency and flexibility for growing applications.
 disadvantages
1.Limited Language Support: Amazon Lex supports a restricted set of languages,
limiting its use for global or multilingual applications.

2.Complexity for Advanced Use Cases: Creating sophisticated conversational

flows and handling complex interactions can be challenging and may require deep
technical expertise.

3.Cost Can Increase with High Usage: The pay-as-you-go pricing model may lead
to unexpectedly high costs as usage scales, especially with large volumes of
interactions.
Microsoft BotFramwork
 1. Define the bot’s Purpose and
Scope
2. Design the conversational Flow
3. Environment Set up
Methodology 4. Create Knowledge base
5. Activity handler chart
6. Create methods for activity
7. Test the bot
8. Deploy the bot
9. Monitor and optimize
10.Update as needed
 Define the bot’s Purpose and
Scope
1. Purpose of bot:
a. To build user-friendly chatbot that
could help to answer common
queries.

2. Scope:
a. Records
b. FQA
c. Tournament intratas
d. Business Information
Design the conversational Flow
For each intent we need to design conversational flow.

a. Entity requirement fulfilment

b. Redirection
c. Context management
Environment Set up
Environment set up is consist of various steps:
Download bot farmwork emulator
Create Microsoft Azur account
Create LUIS account
Create QnA account
Install Yeoman
Install Yeoman
Create echobot
Create knowledge base
Upload all the available knowledge into knowledge base in
the supported formate.

a. pdf
b. xxls
c. url
d. plain text
e. tab seprated file
Activity handler chart

An “Activity” is the basic communication type for the Bot Framework 3.0 protocol.

Activity handler chart is visual representation how a particular conversation will be

initiated and ended describing complete flow including different platforms and
database
Create Methods for activity
handler
We would like to take some action based on activity thus those actions are taken
under the method.

Each activity has a method inside that method we can write our custome code to
handle that activity.
Entity and Intent
Provide all entity and intent to bot in LUIS platform
Test your bot

After development we need to test the bot.

Is Integration between various Microsoft’s platform successful?
How accurately is answering to all questions?
Is user-friendly ?
Monitor and Optimise

Monitoring and optimizing deals with the fact that how we can make our chatbot
more accurate and building it in a way that it answer to the point to the user’s
query.

Along with that we will also try to make it easy to use by removing extra
information that was needed. We will optimize it.
Update as needed

With time new knowledge should be provided to knowledge base.

By analysing user’s query we can make changes into FQA.

Worked on unfulfilled queries.

Microsoft’s Bot Farmwork
Methodology
Advantages
LUIS is highly customizable but may require more setup and understanding of how to tune the model
to suit your needs.
Advantages
Microsoft’s Bot Farmwork
Methodology
Disadvantages
It is used for more advanced use cases and if we have a team already skilled in Microsoft technologies,
Microsoft Bot Framework.
Advantages
Have to pay for all the service we use such as LUIS, Azure Bot Services, and Azure Functions. It’s
cheap if we are already using Microsoft azure.

Disadvantages
Doesn’t have voice first capabilities.
Compariso
n
Factor Google Amazon Microsoft

Ease of Use User-friendly, intuitive Easy to use, integrates Requires more setup and
console with a visual seamlessly with AWS coding, but offers Bot
interface for building services. Framework Composer for low-
bots. code development.
NLP Strong NLP powered by Good NLP, built on Powerful NLP through LUIS
Google’s machine AWS's Lex (similar to (Language Understanding
learning models. Alexa). Intelligent Service).
Can scale, but Easily scalable through Highly scalable via Azure
challenging to manage AWS infrastructure. services (can handle large-
Scalability
many intents and scale, complex bots).
entities.
Customization Limited customization for Supports custom code Highly customizable with
advanced use cases via AWS Lambda, but rich SDKs and integrates
(relies on Google Cloud can be restrictive. well with Azure and external
features). APIs.
Factor Google Amazon Microsoft

Voice Capabilities Built-in integration with Strong voice capabilities Supports speech-to-text and
Google Assistant (good with Alexa integration. text-to-speech via Azure
for voice bots). Cognitive Services, but
requires setup.
Integration with Other Easily integrates with Deep integration with Integrates well with Azure
Services Google Cloud services AWS services like services, enterprise tools, and
like Firebase, Google Lambda, DynamoDB, other Microsoft products.
Sheets. and S3.
Basic role-based access Limited collaboration, Supports collaboration with
Collaboration and control, but limited but integrates well with Azure DevOps for CI/CD, Bot
Team Features collaboration tools. AWS tools for team Framework Composer for non-
management. developers.
Version Control No built-in version No native version Integrated with GitHub and
control, but you can control; relies on external Azure Repos for version
export and import tools like GitHub. control.
configurations.
Summery
Google Dialogflow:

Best for: Developers who need advanced Natural Language Understanding (NLU) and want to
build
multilingual chatbots with rich conversational capabilities.

Ideal Use Cases: Customer support bots, voice and text-based assistants, FAQ bots, and bots
requiring
complex NLP.

Amazon Lex:

Best for: Developers seeking a seamless AWS integration for both voice and text-based chatbots,
especially if you’re already using other AWS services like Lambda or Polly.

Ideal Use Cases: Voice assistants (e.g., Alexa skills), customer service bots, transactional bots, and
scalable chatbots in environments where AWS is the core infrastructure.

Microsoft Bot Framework:

Best for: Teams building cross-platform bots with deep integration into the Microsoft
ecosystem, like Azure,Office 365, and Teams.

Ideal Use Cases: Enterprise-level bots, cross-platform bots (web, mobile, messaging), and bots that
require integration with Microsoft's suite of tools or advanced AI capabilities (e.g., Cognitive
Services).
We have disucssed 3 approch