Smart Chatbot

Introduction

With a combination of artificial intelligence and natural language processing we have developed a Classpie Chatbot. For having a human style conversation through text or auditory methods, this Classpie chatbot serves the practical purpose of customer service and in answering the questions asked by the user. Being power-driven by Google Platform Dialogflow it assists all the teaching institutes for numerous tasks and other activities to handle in a proper way.

Statement of Problem

For interacting and understanding the problem of user and converting it into structured data machine learning proves to be a worthwhile asset. With the leverage of Dialogflow, Classpie chatbot can directly interact with the user and gives the relevant answers according to the utterances by triggering out the intent.

What is Dialogflow?

DialogFlow (called as API.AI also) is a conversational agent building platform on Google. It is a web and application-based platform that can be accessed in any web browser. It enables the developers to create the conversation and handle the tasks of understanding the natural language utterances.


Objective for Creating Classpie Chatbot

We have created a Classpie chatbot that is a virtual assistant tool for having a conversation in the natural language spoken by humans. It is basically developed for the educational domain and specifically for teaching and non-teaching staff section.

For giving you a feel that how a chatbot conversation is all about, we have shared the image to make you understand better:


Intent

An intent is a mapping between the user and action that needs to be taken by software. Firstly, main intent has to be created by the software that matches the user’s order such as “list of students with unpaid fees”. Therefore, the tag here is “unpaid fees” and for the custom entity, we have developed @unpaid.


Creating the Entities

There are basically three types of entities used in dialog flow. The word entities brief us about the extraction of parameter values from the natural language written by the user.

We can create three types of entities:

I. System: Defined by Dialog flow

II. Developer: Defined by developer

III. User: Individual end-user for every request

From the above screenshot of Classpie chatbot we have fetched the intents. With the assistance of image shown below, you will exactly get to know about what are entities in the dialog flow:


Action and Parameters

I. Action

When a specific intent has been input from the user, then the action step is being taken up. From the users request the information is extracted and it appears in the following format:

{“action”: “unpaid”}

II. Parameters

These are the words for connecting in a user’s response. In response to any query from the user the parameters are returned in the following format:

{“parameter”: “branch”}


Results

The results can be download in the PDF or excel sheet. Adjacently, the users can view them on rich web UI also. Here, we have shared the image of results that will be displayed to a user.


Plan of Action

I. Natural Language Processing (NLP): With the leverage of dialog flow the users can communicate with the Classpie Chatbot in the natural sentences. All the input sentences typed by the user are matched by dialog flow. With the command function, the intents are fetched up and predetermined responses are delivered to the users.

II. Understanding User Intent: For mapping synonyms with the programming entities Dialogflow is being used. The platform is already trained with major entities and intents that can describe activities in a teaching institute. Google's speech to text understands the question correctly and maps to the right intents. Based on this an SQL query is generated that gets executed on MySQL database containing all information of a teaching organization. After it, the user’s intent is being matched and send for an action.

III. Machine Learning: There are many input statements from the user for which the bot doesn’t respond. By analyzing conversation logs the intents can be improved and the bot will be trained to respond for the questions asked by the end user.

Benefits of Accessing Classpie Chatbot:

There are multiple number of benefits of accessing a Classpie Chatbot. Here, we have discussed a few of them.

I. For Gaining Insights: For the communication, tracking and monitoring the activities of students and teachers in an educational institute. It eases the work of upper management staff and authorities to fetch up the data. While monitoring the data, better strategies and plans can be organized for efficiently and well-organized working in a streamlined way.

II. 24×7 Support: With extensive assistance, the user can proactively interact and quickly get enriched content in the form of videos, images, text etc. The Classpie chatbot provides real-time assistance by interacting with the representative and understanding their real problem in their natural language.