Fourth International Conference on Electronics, Communication and Aerospace Technology (ICECA-2020)

IEEE Xplore Part Number: CFP20J88-ART; ISBN: 978-1-7281-6387-1

Dynamic Information Retrieval With Chatbots: A

Review of Artificial Intelligence Methodology

Vishal Dutt Satya Murthy Sasubilli Anand Eswararao Yerrapati

Department of Computer Science, Workday Integration Architect, Solution Architect
Aryabhatta College, Huntington, Columbus, Oracle,
Ajmer, India OHIO Bangalore, India
vishaldutt53@gmail.com satya.Murthy@huntington.com anand.yerrapati@gmail.com

Abstract— A chatbot is a computer program designed to different channels and messenger apps. Because of this, there
simulate a conversation with human users, especially over the is no need to develop multiple chatbots for various channels.
Internet. Ultimately, the chatbot acts like a virtual assistant or So, any improvements to the chatbot can be done to one
interactive agent in a conversations interface to respond to user version, where it would apply to all the channel users.
queries or messages via communication channel like mobile apps,
messenger apps or browser-based applications. Chatbots have In this paper usage of chatbots in HR domain is
become more popular nowadays and most of the companies are highlighted and designed to provide a solution for job search
implementing them wherever they can to reduce the operation and job apply in different formats.
cost. In many cases, human resources are utilized to respond to
user queries, where the chatbot can do the same job by searching Natural Language Processing (NLP) is the great tool to
the data in the system so that the human talent can be used for work on understanding the human inputs. The sentimental
other advanced tasks. As the advancement in the technology, analysis of the human emotions is the key to design a model
chatbots are also evaluated in a better way such that they can do which can work on chat bots which can help the human
some other tasks beyond just answering the textual questions. resource department. The recruitment with the technology
This paper provides a chatbot based solution for users or increases the production of the great team. Artificial
candidates, who are searching for a job to apply in a company. intelligence is the superior basement for development of the
This solution makes the job searching and applying process easy, great things related to the real-time problems. The natural
where the user can apply for a job in a few taps without visiting language process will work on stemming and lemmatization
the company website or their mobile app. which can further enhance the performance to model which
was tried to build[2].
Keywords— Chatbots, Artificial Intelligence, Recruitment, Job
Search, Job Application, Mobility The process of stemming in natural language processing is
to understand the concepts related to the grammatical things
related to the language the user has chosen. In chat bot
I. INT RODUCT ION
creation we can’t restrict the user to follow some specific
It is very rare to find people, who are not using mobile and language to operate and comminate with the machine. The
a messenger. With the availability of internet and increase in communication with device will be based on the type of
the usage of mobiles, almost everything is available at the language and phrases, while designing the concept. The
doorsteps. Then why can’t be jobs? Even today if it is required concept behind using the natural language process ing [NLP]
to find and apply for a job in a company, either the candidate model is to understand the query generated by the user or the
should visit the company website or install their mobile app. person, who is attending interview. The interviews , which
Sometimes, applying job in website might be tedious. The have been done by the person in physical can be replaced by
process always require a laptop or desktop to fill all the the machines in the secured channel, where it can extract the
required input fields. Why can’t the mobile messenger apps knowledge on different people, who is having different
are used to apply for the same jobs in a company? domain knowledge. The different domain knowledge is
Chatbots provide a solution by connecting to any of the important to hire a person for different roles in the
messenger apps like Facebook, Twitter, WeChat, Line, Slack organization[3-4].
etc. So that there is no need to install any mobile app of the Stemming and lemmatization are the common strategies to
company just to search and apply jobs. As the advancements handle the multiple queries. For an instance we can consider
in the chatbots, they also support multiple languages, which is the questionaries’ posed by the user to the learning platform
very much helpful for localization. And the effort would be support team and they requested for some answers for their
less. Many chatbot platforms support multiple channel [1] questions using the bot and that bot was programmed statically
integrations. Hence, the same chatbot can be exposed to because it can’t take the inputs from the user in dynamic

978-1-7281-6387-1/20/$31.00 ©2020 IEEE 1299

Authorized licensed use limited to: Cornell University Library. Downloaded on May 23,2021 at 22:05:43 UTC from IEEE Xplore. Restrictions apply.
 Fourth International Conference on Electronics, Communication and Aerospace Technology (ICECA-2020)
IEEE Xplore Part Number: CFP20J88-ART; ISBN: 978-1-7281-6387-1

format. If the user gave the input in the format prescribed in sentimental analysis of the human inputs which can carry th e
the form of syntax, then the machine can understand the all the aspects of the requirements mentioned by the user. The
phrase mentioned and process the query. user will give the sample inputs and make the application to
understand the requirement. The requirement may be based on
Rest of the paper is organized as follows. Section II the experience as well the domain of the work. There is a
discuss the use case and requirements. Section III describes
simulation which can make us to understand the requirement.
the design and proposed solution for this paper. Section IV
shows some screen shots of the interaction with the designed If some of the online job portals are observed, when the
chatbot solution and finally Section V concludes the paper. profile is updated as a fresher, they will show the jobs with
one year experience, if it is updated with one year, the job
II. USECASE DISCUSSION results will show with the experience of the 2+. In this way the
people who seeking the job change are not getting the right
Here, in the use case, user should have an easiness to find offers.
and apply to a job in the company. A chatbot is required to
interact with the user like a human to ask and understand the The real time applications with chat bot can solve the
user requirements and suggest the user with best available job. problem as many researchers thought the same. Well that’s not
The bot should be flexible to take the resume in multiple that easy to understand the machine which can track the
formats (pdf or doc or any file) or a form with input fields. information related to the actual user inputs. The inputs given
And also, the bot should give more information about the to the machine will be tracked for the further reference. This
company. kind of reference will make you to understand the requirement
of the users and chat bot will give the bes t out of the
The main purpose of the bot was divided into different condition.
parts and those are mentioned as follows:
x First the bot needs to understand the language the IV. EXISTING APPROACHES
user communication with it There are different existing approaches , which are
x The bot needs to understand the key things in the speaking about the bot, where it is performing job-based
question it has to ask. analysis and service. The performance of the bot will be based
on the user criteria. The purpose of the bot needs to be specific
x Need to get the company related answers and dynamic.
x Need to understand the different reviews given on
that specific role in the company
Several existing methodologies are working on the chat
x Need to understand the specific review on the bot. In early stage of artificial intelligence, a medical chat bot
company before asking them to answer questions can be used to perform the task like understanding the human
symptoms by different questions. The questioning process will
x The bot needs to be dynamic in understanding the be based on the symptoms input we gave to the machine it will
answers even though there are spelling mistakes identify the disease[9].
The main purpose of the chat bot is to understand the
person in a dynamic format and the user needs to understand
the exact requirement that a bot post to the user. There is another methodology in job identification
mechanism by using the chat bot in recent times , which is
III. LITERATURE REVIEW asking about the experience in a specific field and trying to get
the related postings on the platform to the user. But it fails to
The chat bots are most prominent technologies which can understand what is there in the resume and whether it is
take the review regarding the user input and then they process matching with the inputs given by the user or not. The
it using the machine learning models , which can considerthe comparison between the resume and the contents mentioned in
process model to predict some information from that data. The the profiles must be performed for the better understanding of
process of managing the job requirements using the chat-bots the requirements of the user.
is a challenging issue, where it is required to focus on the
latest technologies, which can implement the human nature Some of the applications of chat bot are trying to force
identification. There are so many ongoing researches on the the user to choose the contents only mentioned by the owner
concept of artificial intelligence and made a way beyond of the application, but the search results and the operations
thinking for the real time problems. The main real time will not be based on the requirement of the application. The
problem which the work mentioned in the proposed system is requirements must match with the concepts mentioned in the
regarding the method of implementing AI into chat bot, where resume. The same thing will apply in all aspects of analyzing
it can track the accurate job for the user[5-6]. the content.

The recent times we can find different applications which Different applications are there and real-time
are working on the same thing like chat bot which can take the applications on internet is having this kind of chat bots. For
user inputs and make them to work on understanding the example, the COVID – 19 application is considered, where it
human requirements. The requirement is to understand the asks the symptoms of the user and identifying whether the

978-1-7281-6387-1/20/$31.00 ©2020 IEEE 1300

Authorized licensed use limited to: Cornell University Library. Downloaded on May 23,2021 at 22:05:43 UTC from IEEE Xplore. Restrictions apply.
 Fourth International Conference on Electronics, Communication and Aerospace Technology (ICECA-2020)
IEEE Xplore Part Number: CFP20J88-ART; ISBN: 978-1-7281-6387-1

patient is having the virus or not. In the same way some chat bot’s job is not only giving the result for the searched
applications are already available but the major difference is query but also to understand the important things related to an
when there will be difference in the title of the job or the query organization and also related to the company. The main
and the required answer. There will be rapid difference in all agenda here is to understand the user inputs in the form of
these things. It is required to process the information based on reviews. If the review of a certain thing is not given properly,
the accurate requirements and in this position, it requires then the application developer need to look over the co ncept
machine learning methodologies [8]. of language processing and need to compare the real quality of
the product and the review given by the user on the same
V. PROPOSED SOLUT ION product. Some of the competitors will make a fake review on
In this work, we developed an interactive chatbot for a the genuine product or the application, then we need to make
company SonuAppz (This name is taken only to understand the model to understand such kind of issues. In the same way,
the work more clearly) and the workflow of the proposed if any employee was hired from the organization and gave
solution is shown in Fig 1 and in section V. The flow starts some bad report intentionally on the open platform, then it
with the user interaction. The user can type any message, let's may create the problem monitor the exact meaning of the
say "Hi". If the text is not a formal question, then the bot application. The next section will deal with the result analysis
responds with the available options like menu it handles as of the application developed [14].
shown in Fig 2. The user can tap on the option "Explore with Different applications are available on the same kind of
me" before searching for a job to know more about the concepts, like application development using chat bot and
company as shown in Fig 3. This option helps the user to managing the organization activities with chat bot. but there is
understand the work culture and company benefits like leaves, a situation where chat bots are failed to develop the user-
policies, growth options in the company etc. Now the user can friendly environment for the users. Even the search engine
select the "Search Job" option and the bot responds as shown fails to give the accurate result some times. The proposed
in Fig 4. system will separate the positive and negative cases in the user
If you observe Fig 4, the user can find jobs by entering a job input and check the compatibility with the query available in
type like Sales, Marketing or he/she can find by location by the repository and compute the operation [15].
providing his/her location. To get the location, click on the Searching for the job is not a simple task. We need to
button "Based on location" which will open the map and understand the user requirement in and out. The process of
allows the user to select a location. If the user doesn't find any implementing the proposed approach satisfies all th e
jobs with these two options or he/she is not clear wh ich job requirements based on the users input justification with both
fits him/her, then he/she can select the option "Assist me". positive and negative instances.
When the user selects this option, the bot will ask a few NLP is the common approach of machine learning to design
questions like his current organization, job category etc. and the chatbot process. The NLP will work based on the corpus
suggest the relevant jobs if any[10-11]. and the reality will be based on the user inputs. Corpus is
The application developed working on different approaches on having the positive and negative words. But NLP the proposed
the single entry. When the user enters into the chat section it method working on is to analyze and reduce the negativity in
will ask for the different things being provided by the user search.
application provider or the service provider. Then it must The query the user posed will be compared to the existing
activate the segment that query belongs to. For an instance openings in the repository and there will be comparison
consider that the user requires some answer to the important portion using NLP which will identify the accurate keywords
question, which is not related to the job. For an instance if the of the search should the user need to ask and search process
user requires some information related to the acceptable notice will be refined internally.
period of some organization, need to check with FAQ section.
VI. RESULTS
After selecting that section, it will show the pre-defined
questions for the user to answer those. Then it will check the Chatbot is implemented to meet the user requirements to find
information requested by the user and process it as the relevant jobs and information of an Organization. This section
important query and make it valid after validating the user shows some screenshots of the developed chatbot, which help
inputs. The architecture following here is in three ways. But it to understand this paper better.
will connect to the specific DB according to the selection. As Rather than the content mentioned in the previous section, we
mentioned earlier, if you choose FAQ, then it will connect to also need to look over the following results of the application
the FAQ DB then process the requests further. In each case we SONU APPZ. This application can be worked with different
have individual DB’s for the processing and the requests are approaches and priorities. This application was developed to
processed according to each step we have chosen[13]. make the user to get the better job chances and the application
The NLP will work on the different methodologies. If the will give the accurate responses to the users to get the accurate
request is regarding the job assistance, then it need to result.
understand the concept behind ques tion posed by the bot to the This kind of applications are not developed by using the
user and it has to understand the insights of the resume. The filters. This is an artificial intelligence application which can
chat bots must work on the sentimental analysis. Because, a hold the values.

978-1-7281-6387-1/20/$31.00 ©2020 IEEE 1301

Authorized licensed use limited to: Cornell University Library. Downloaded on May 23,2021 at 22:05:43 UTC from IEEE Xplore. Restrictions apply.
Fourth International Conference on Electronics, Communication and Aerospace Technology (ICECA-2020)
IEEE Xplore Part Number: CFP20J88-ART; ISBN: 978-1-7281-6387-1

Fig. 1. Chatbot Workflow

Fig 4: Chatbot response for Search jobs

VII. CONCLUSION
The chat bot applications are most required thing in the
development of an organization. In the current application,
there is an interaction of the user and a model to understand
the insights of the user inputs in the form of question and
answer. Natural language processing is the major mechanism
adopted in this application. Lemmatization and stemming are
the key concepts to discuss about the process of having a
perfect chat bot. In this application, the user needs to select the
type of communication from the service provider, then it
performs the operations like data manipulation, and data
addition in the application. Everything is based on the user
inputs based on the selection. This application is mostly used
for the better implementation of the job-oriented platform by
using chat bots instead of struggling for the job search.

References
Fig. 2. Features the bot offers to user
[1] G. Eason, B. Noble, and I. N. Sneddon, “ On certain integrals of
Lipschitz-Hankel type involving products of Bessel functions,” Phil.
T rans. Roy. Soc. London, vol. A247, pp. 529 –551, April 1955.
(references)
[2] J. Clerk Maxwell, A T reatise on Electricity and Magnetism, 3rd ed., vol.
2. Oxford: Clarendon, 1892, pp.68–73.
[3] I. S. Jacobs and C. P. Bean, “ Fine particles, thin films and exchange
anisotropy,” in Magnetism, vol. III, G. T . Rado and H. Suhl, Eds. New
York: Academic, 1963, pp. 271–350.
[4] K. Elissa, “ T itle of paper if known,” unpublished.
[5] R. Nicole, “ T itle of paper with only first word capitalized,” J. Name
Stand. Abbrev., in press.
[6] Y. Yorozu, M. Hirano, K. Oka, and Y. T agawa, “Electron spectroscopy
studies on magneto-optical media and plastic substrate interface,” IEEE
T ransl. J. Magn. Japan, vol. 2, pp. 740–741, August 1987 [Digests 9th
Annual Conf. Magnetics Japan, p. 301, 1982].
[7] M. Young, T he T echnical Writer’s Handbook. Mill Valley, CA:
University Science, 1989.
[8] “ Anatomay and utility of an artificial intelligence conversational
entity”, Anirudh Khanna et al, 2015, IEEE conference
[9] Evaluation of authorship attribution software on a chat bot corpus,
Nawaf Ali et al, 2011, IEEE
[10] “ Absolute Answerability In T he Era Of Artificial Intelligence And
Machine Learning: A T alent Management Perspective”, Preethi Keerthi
Dsouza, 2019, IEEE.
Fig. 3. Options to know about the company
[11] “ Intelligent Chatting Service Using AIML”, G. Saqib et al, 2018, IEEE

978-1-7281-6387-1/20/$31.00 ©2020 IEEE 1302

Authorized licensed use limited to: Cornell University Library. Downloaded on May 23,2021 at 22:05:43 UTC from IEEE Xplore. Restrictions apply.
 Fourth International Conference on Electronics, Communication and Aerospace Technology (ICECA-2020)
IEEE Xplore Part Number: CFP20J88-ART; ISBN: 978-1-7281-6387-1

[12] “AI Based Chat -Bot Using Azure Cognitive Services”, Kapil T ajane et [15] “ DocChat: An Information Retrieval Approach for Chatbot Engines
al, 2019, IEEE international conference Using Unstructured Documents”, Z. Yan et.al, 2016, Proceedings of the
[13] “Artificial Intelligence Based Solarbot”, Sayali Patil et al, 2018, IEEE 54th Annual Meeting of the Association for Computational Linguistics
Volume 1, DOI: 10.18653/v1/P16-1049.
[14] “HR Based Interactive Chat bot (PowerBot)”, Shabana tadvi et al.,
2020, IEEE

978-1-7281-6387-1/20/$31.00 ©2020 IEEE 1303

Authorized licensed use limited to: Cornell University Library. Downloaded on May 23,2021 at 22:05:43 UTC from IEEE Xplore. Restrictions apply.