![The basic structure of Bershca [18]](https://www.researchgate.net/publication/369876814/figure/fig1/AS:11431281139724456@1680887397327/The-basic-structure-of-Bershca-18_Q320.jpg)
Available via license: CC BY-NC 4.0
Content may be subject to copyright.
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
339
Chatbot Improves the Customer Service in Four Important
Industries
Xin Zhou
Donald Bren School of Information & Computer Science, University of California, Irvine, CA, USA
Xinz36@uci.edu
Abstract. Chatbot was first introduced in 1966. With the development of Artificial Intelligence for
decades, the function and usage of Chatbot have been enhanced to a very high level. In this paper,
we are going to talk about how the use of Chatbot improves the customer service in four industries:
healthcare, advisory, commercial, and education. They are the four important social fields, and they
all have very large demands and uncountable customers every day. The human agents are facing
tons of customers from various backgrounds. They need the help from artificial intelligence. Apart
from that, we will also discuss some special advantages of AI-based chatbots which them work so
effectively in customer service. These three advantages are automation, trust, and productivity. With
the combination of these advantages, chatbot can be highly supportive in the field of customer
service and improve the quality of customer service. Eventually, we will talk about some challenges
the chatbots are facing now, and how to make it function better in customer service to meet more
users’ expectations.
Keywords: Chatbot; Trust; Automation; Productivity; Customer Service; Education; Commerce;
Healthcare; Advisory.
1. Introduction
In 1950, Alan Turing, who was considered as one of the greatest computer scientists in the history,
proposed an idea “Can machines think?” Then, in 1966, the first chatbot named ELIZE was created
by MIT, and it was able to generate satisfactory response according to keywords and matched patterns
[1]. Now, in 2022, Artificial Intelligence is considered as one of the most important and advanced
technology in the world. The improvement of computer technology has brought chatbots with much
stronger abilities. The reason of developing better chatbots is linked with the large demand. Why
people need chatbot?
In order to figure out why people need chatbots in their daily life, we need to find out what
advantages chatbots bring to people. There are mainly four advantages – ease, speed, convenience,
and quick access to information [2]. When we are talking to a chatbot, all we need to do is to type the
query on the user interface. Then, the chatbot will generate the response according to our inputs in a
very short time. The whole process is very easy and quick. Also, most of chatbots are accessible
because they are used online. As long as our mobile devices are connected to Internet, we could talk
to various chatbot according to our own needs. Some people might ask why the users prefer talking
with a computer-based agent instead a real human agent. One study was conducted to compare
random human IM, which is Instant Messaging, with random human conversation with chatbot to see
the difference between human-to-human. The result shows that human beings send less words to
Chatbot than to real human. Also, human beings are modeling their communication with chatbots,
like the way they are adapting their own language [3].
Next, we move our sight to a field which seems far away from computer science – customer service.
Customer service is defined as a management strategy which is used to meet the customer’s
expectations, and the agents working in the business need to understand what the customers need in
order to provide satisfactory responses [4]. According to this definition, it seems like we can build
some connections between customer service and chatbot because chatbots are able to generate
responses according to the users’ input.
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
340
Then, let’s specifically talk about what chatbot’s functionalities could be used in the field of
customer service. Generally, chatbot can be split into five types: task-oriented, social, service-based,
knowledge-based, and dialogue-based. The dialogue-based chatbot can be used in customer service
because it is designed to engage with user. It is composed with three parts: User Interface, Data
Processing, and Integration Components [5]. With these functionalities, and some advantages of
chatbot we talked in paragraphs above, applying chatbots in the field of customer service is very
possible. Compared to real human agents, computer-based chatbot is able to provide service in 24
hours and 7 days each week. However, only with these qualities, we can’t conclude that chatbot can
improve the customer service. Hence, we are going to talk about more in the following paper.
Customer service is needed in many industries. It is impossible to talk about the application of
chatbot in each of them. Finally, we select four of the most important industries that are linked with
all the people in society. They are Healthcare, Advisory, Commercial, and Education. Through lots
of research about the connection between chatbot and customer service online, we conclude that
chatbot is able to improve customer service in these four industries.
2. Advantages of Chatbots
2.1 Automation
Automation can be considered as the most important property of chatbots. The chatbot can be
defined as “automate system that emulates person-to-person dialogue through text or voice message
[6].” According to this definition, we can see that the original purpose of designing and inventing
chatbot is to finish automated tasks. Then, if we move our sight to the field of customer service, we
can see that it really needs the help of chatbot. The customer service continuously has more
improvements and larger extension in recent years. As a result, there are more and more customer
requests [7]. The human agents are not able to provide help to all the customers at the same time.
However, with the help of chatbot, this problem can be solved very easily because of chatbot’s
automation ability. Chatbot is able to talk to many users at the same time and provide personal service
to each customer who needs help urgently.
2.2 Trust
The word “trust” is considered as having faith on a thing or a person. It means that if we trust
someone, we have faith on him/her. Then, how are we supposed to put truth on machine. When we
are talking with a chatbot, we need to share some personal information with it so that the chatbot will
be able to generate useful responses for us. Then, there is a concern about this part – the privacy of
our personal information. A survey conducted by Saglam, Nurse, and Hodges on the SurveyMonkey
platform focused on the potential risks concerning interaction with chatbot and factors which make
them have trust on chatbot. According to the result of their survey which is on Fig. 1, we can see that
the most two concerns are “how to delete” and “inappropriately used” of their personal data [11].
What if our own data is not deleted after the end of conversation? What if our own data is
inappropriately used by the company behind the chatbot? Apart from that, this survey also asked the
participants about why they have trust on chatbot. The top two responses are “response quality” and
“grammatical correctness [11].” We can that the quality of response and the context of response really
matter. Also, another survey conducted by Følstad, Nordheim, and Bjørkli found more specific factor
that make users have trust on Chatbots. Those factors are “interpretation and advice”, “human-
likeness”, “self-presentation”, and “professional appearance [12].” “Interpretation and advice” refer
to the notion of the qualities of machine interpretation and response. “Human-likeness” is how similar
is the chatbot talking like a human. “Self-presentation” is related to the chatbot’s communication skill,
and “professional appearance” is determined by how professional the chatbot is in the specific field
[12]. The uses have more trust on chatbot if the chatbot can do well on those four factors we talked
above. However, these are all the user-related factors. There are also environmental related factors.
Another paper, also wrote by Følstad, Nordheim, and Bjørkli, explained two environmental factors
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
341
that determine the users’ trust on chatbot. The first one is risk which is the undesirable outcome by
talking with the chatbot. The second factor is brand. Brand is the service provider, and it can be
considered as the company behind the chatbot [7]. If the service provider is well-known company,
like Microsoft and Google, the users might have more trust on its chatbot.
Fig 1. User’s concern of using chatbots [11]
2.3 Productivity
The productivity of chatbot can be split into two parts: productivity to users and productivity to
the company behind the chatbot. First of all, let’s talk about how chatbot improves the users’
productivity. It is saying that the users expect the chatbot to be effective and efficient [8]. The only
thing we need to do is to tell the chatbots what we need, and then it will search the database that
match the input query the best. The whole process is very easy, and it is very fast so that it saves the
users a lot of time. Also, with the improvement of Machine Learning and the using of better algorithm,
the chatbots are able to find generate the satisfactory response. Effectiveness and efficiency relate to
productivity. In most cases, the motivation for the customers to use chatbot is productivity [9]. The
users are enjoying the productivity brought by chatbot. Apart from that, the chatbot is able to improve
the productivity to the company behind. In business, the use of chatbot makes the social and
organization relations more easier and user friendly [10]. Some companies are designing the chatbot
for employee management. Chatbots can help with tasks like employee record maintenance and
performance review [10]. With the help of chatbot, these tasks are turning to be much easier. Then,
let’s move our sight to the company which needs to provide satisfactory customer services. Currently,
by using the chatbot, the customer service has been improved to a new level. One important
characteristic of chatbot is automation. It is almost impossible for one human agent to talk with
several customers and help them to solve the problems. However, chatbots can do it easily. The
chatbots are able to talk to each customer and generate responses according to the user’s query. Also,
the study finds that people are willing to have conversation with chatbots [3]. The chatbots help the
company to deal with the problem of overwhelming customer service and increase the company’s
productivity.
3. Application in Four Industries
3.1 Commercial
Commercial is a very big industry in our society. By talking about the application of chatbots in
this industry, we mainly focus on e-commerce. E-commerce is defined as “the use of e lectronic means
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
342
and technologies to conduct commerce [14].” Most of chatbots used in e-commerce industry are task-
oriented [13]. Compare to human agents or human sellers, what advantages do chatbots have?
There are actually a few advantages of using chatbots in e-commerce. The first one is cost.
Applying chatbots has a lower cost than human representatives, and chatbots only need maintenance
in a certain period. Compare to the human agent’s monthly salary, the chatbots’ maintenance cost is
much lower. The second advantage is that chatbot makes less mistake. Since chatbots generate
responses based on its knowledge, the answers are more reliable [15]. It’s a very important factor of
customer service in e-commerce to generate correct answers to make customers have trust on products.
Also, compare to human agents, AI-based chatbots won’t have negative emotions or feel exhausted,
and they are always positive [16]. After one day’s working, the human agents might feel tired, and
it’s hard for them to always keep a positive emotion while facing the customers. Now, with the
application of chatbots in e-commerce, they are able to provide positive attitudes towards the
customers all the time. It could enhance the customer service. The third advantage of chatbot is its
ability of handling large number of customer communications [16]. The purchases of e-commerce
happened online which means that the chats between the customers and agents are through typing or
oral talking. The human agents might have the problem of having several customers at the same time,
but they can not provide responses at the same time. However, this work can be done by chatbots
easily because they can talk to several customers at the same time and generate responses according
to each customer’s demand. Then, the customers won’t have the problem of waiting in line.
Sometimes stuff are asked with the same question by a variety of customers, but chatbots make it
more cost-effective [17].
There is a chatbot named SuperAgent. It has three advantages. First, it can leverage crowd-
sourcing styles. Second, it uses NLP (Natural Language Processing) and machine learning techniques.
Third, it integrates into e-commerce websites as an add-on extension [17]. With these three factors,
SuperAgent is able to make customers’ shopping experience better.
Actually, apart from e-commerce industry. Chatbots can be used in other commercial places like
hotel. Gunawan et.al introduced a chatbot named Bershca used in Indonesia. It has shown that the
customer satisfaction has great impact to hotel’s profit. This chatbot is a front-end application. The
front part is built by Google Flutter, and the end part is based on Python plus AIML (Artificial
Intelligence Markup Language) [18]. The basic structure of Bershca is on Fig. 2.
Fig 2. The basic structure of Bershca [18]
3.2 Healthcare
Healthcare is so important to us because we get illness, and we need treatments for our health. In
the past three years, COVID-19 has become a global virus. Everyday there are more people infected.
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
343
However, the face-to-face service is limited to prevent the spread of virus so that there’s more need
of digital self-service to help customers [19]. Chatbot, which is the AI-based technology, has the
ability to talk to several customers and generate response according to the customer’s needs. Hence,
chatbots are adopted by lots of hospitals to provide healthcare. For instance, some famous healthcare
institutions like Babylon Health, Ada Health, and Your.MD offer chatbots to check patients’
symptom [20].
Why we need chatbots in healthcare and how they can help patients? Sometimes, when the patients
are going to hospitals, it is time-consuming for some people. In this situation, they can reach to
chatbots for help because chatbots provide users with real-time interactive service at any time and
any space [21]. This is one advantage of chatbot in healthcare because the hospital might have patients
at any time. Also, Winkler and Söllner asserts that chatbots are able to provide patients with
information and consultations at the time of hospitals discharge [22]. Additionally, now chatbots have
the ability of recognizing users’ emotions and generate response based on the current patients’ mood.
Chatbots can show empathy to patients [23]. It helps patients to relax. It turns out that patients with
depressive symptoms prefer to receive help and support from chatbots [23]. Another advantage of
chatbot is that it has the potential to improve access to health care by providing remote counsel with
low cost. It helps some patients with medical burden [20].
3.3 Advisory
As the name “advisory” implies, the main goal of this industry is to provide customers with advice.
Sometimes people need advice from other people because they are not able to make a decision by
themselves. In order to improve the customer service, the advisory chatbot is invented. The role of
this type of chatbot is to provide suggestions and give recommendations on service. Also, it can
contact people and provide users with support and advice [24]. In this section, we are going to talk
about the advisory chatbot used to provide career choice and advisory chatbot used in school.
In lots of cases, the unemployment rate is connected with the lack of proper career guidance [25].
People lose their jobs because they choose the wrong jobs which do not fit for themselves. In order
to avoid this problem, the advisory industry needs to provide people with the most suitable jobs
according to their backgrounds. A career counselling chatbot implementation method is introduced
by D’Silva et.al. Basically there are three step. First, they conducted psychometric test on the
participants to identify their personalities. Second, they trained and mentored the participants to find
the suitable job. This step helped chatbot to find the participants’ learning style and know more about
them. Third, the chatbot generated a professional resume according to data gathered from participants
and provided a list of recommended companies [25]. The result generated in this way is more
personalized and suitable. Apart from employed people, students in the college also need career
suggestions because they are going to start working after graduation. The chatbot using Microsoft
Bot Framework is introduced to help to write CV and find jobs [26]. To train the bot and get useful
data from the users, this chatbot is constructed with seven main entities in Fig. 3.
Fig 3. Seven main entities of chatbot [26]
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
344
Students not only need advisory for career choice, but also they need advisory in school. In campus,
sometimes students are having trouble with school’s information, but they can’t always reach to
school’s staff for help. In this situation, chatbots can provide them with assistance. The chatbots used
in school can provide students with a list of helps. 1. Help students to acquire information about
admission date, course cut-off days, and course eligibility. 2. Help students discover classroom and
find building location. 3. Help students to find information on school’s activities and competitions. 4.
Help students with interviews and job preparation [26]. By taking a look at a real example of chatbot
used in school, we can see how much help the chatbot is able to bring to students. There is a chatbot
implemented based on AIML (Artificial Intelligence Markup Language) to help students at Manipal
University. Its main job is to help students to fetch information from school’s system. The information
includes the rank of university, availability of service, university environment, updates of activities,
and academic information [27]. Then, the students do not have to search website for specific
information or go to staff’s office to ask questions. It’s the special customer service for students at
university, and it has been improved by the application of chatbots.
3.4 Education
Education is very important to all the human beings. Education helps us to gain knowledge, have
the ability to interpret other things, make us better citizens, bring us with confidence, ensure a better
future, and helps us to build our own characteristics. (50 From primary school to college, education
is a very long trip. Hence, education is a very important industry in our society. How are we supposed
to link this important industry with chatbot? Actually, chatbots are still at the very beginning stage in
education, but it has proven that chatbots have positive effects on students’ learning process and
student satisfaction while learning [22].
By designing the chatbots for education, we are aiming three goals. 1. Improve communication. 2.
Minimize ambiguity. 3. Increase productivity [28]. Communication is very important in education,
and other studies found that chatbots have great potential in education because of its ability to
communicate through natural language [29]. With the development of machine learning and Natural
Language Processing, chatbots have been improved to talk like a real human being. One important
reason to adopt chatbots in education is because of global health crisis, so there’s more need of
technology methodology, and schools need to implement new information and communication
technologies for remote learning [29]. Chatbots can provide help because they are able to answer
questions online and provide students with help for learning process. It has been shown that chatbots
can help students to learn in a more didactic and efficient way [29]. In the “productivity” section
above, we talked about how chatbot improves the user’s productivity for work. Productivity is also
one important element of study. If students have higher productivity while using chatbots for study,
they would be more efficient.
Chan et.al introduced the audience with a chatbot for elective course selection, which named
EASElective from Open University at Hong Kong. Its main job is to provide students with elective
course advising [30]. Sometimes students have trouble with reaching school staff for advice while
choose the elective course. At this time, students can ask chatbots are help. Chan et.al mentioned that
there’s more diversity within student’s group so that there’s more need of this intelligent chatbot to
provide advice to students [30]. Apart from providing course advises, chatbots can do other tasks like
sending reminders of exams. Also, the chatbot’s generative system can help students to understand
the curriculum. The developing of NLP (Natural Language Processing) helps the chatbot to
understand student’s questions [31].
4. Summary
According to those advantages and studies, we can conclude that the using of chatbots can improve
the customer services. There are three important factors of chatbots which are automation, trust, and
productivity. They make customers have the interest of using chatbots. The satisfaction of using
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
345
chatbots improves the quality of customer service. Commerce, healthcare, education, and advisory
are four important industries in our society which have impacts to almost all the citizens. Nowadays,
chatbots have been used in all four industries. We found that chatbots are able to improve the customer
service in these industries. Currently, chatbots are still at the beginning stage of markets. We can still
make improvements on chatbots. For instance, chatbots do not have memory which means that each
time we are talking with chatbots, they can’t connect the current conversation with the last
conversation. Also, even though we talked about the factor of trust of chatbots, customers can’t have
trust to chatbots all the time because they are worrying about the leak of their own personal
information. All in all, we believe that the future of AI-based chatbots is bright and chatbots could be
applied in more industries.
References
[1] Cahn Jack. CHATBOT: Architecture, Design, & Development. University of Pennsylvania, 2017.
[2] Brandtzaeg P. B., Følstad A. Why People Use Chatbots. SpringerLink International Conference on
Internet Science, 2017, 10673: 377-392.
[3] Hill Jennifer, Ford W. Randolph, Farreras Ingrid G. Real conversations with artificial intelligence: A
comparison between human-human online conversations and human-chatbot conversations. ELSEVIER
Computers in Human Behavior, 2015, 49: 245-250.
[4] Wagenheim George D., John H. Reurink. Customer Service in Public Administration. JSTOR Public
Administration Review, 1991, 51(3): 263-270.
[5] Nuruzzaman Mohammad, Hussain Omar Khadeer. IntelliBot: A Dialogue-based chatbot for the insurance
industry. ELSEVIER Knowledge-Based Systems, 2020, 196.
[6] Um Taehyee, Kim Taetyung, Chung Namho. How does an Intelligence Chatbot Affect Customers
Compared with Self-Service Technology for Sustainable Services? MDPI Economic and Business
Aspects of Sustainability, 2020, 12(12).
[7] Nordheim Cecilie Bertinussen, Følstad Asbjørn, Bjørkli Cato Alexander. An Initial Model of Trust in
Chatbots for Customer Service – Findings from a Questionnaire Study. OXFORD ACADEMIC
Interacting with Computers, 2019, 31(3): 317-335.
[8] Brandtzaeg Petter Bae, Følstad Asbjørn. Chatbots: changing user needs and motivations. ACM
Interactions, 2018, 25(5): 38-43.
[9] Adamopoulou Eleni, Moussiades Lefteris. An Overview of Chatbots Technology. SpringerLink Artificial
Intelligence Applications and Innovations, 2020, 584: 373-383.
[10] Majumder Soumi, Mondal Atreyee. Are chatbots really useful for human resource management?
SpringerLink International Journal of Speech Technology, 2021, 24: 969-977.
[11] Saglam Rahime Belen, Nurse Jason R.C., Hodges Duncan. Privacy Concerns in Chatbot Interactions:
When to Trust and When to Worry. SpringerLink Communications in Computer and Information Science,
2021, 1420: 391-399.
[12] Xu Anbang, Liu Zhe, Guo Yufan, Sinha Vibha, Akkiraju Rama. A New Chatbot for Customer Service on
social media. ACM Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems,
2017, 3506-3510.
[13] Elsholz Ela, Chamberlain Jon, Kruschwitz Udo. Exploring Language Style in Chatbots to Increase
Perceived Product Value and User engagement. ACM Proceedings of the 2019 Conference on Human
Information Interaction and Retrieval, 2019, 301-305.
[14] Manzoor Amir. E-Commerce An Introduction. LAP Publishing, 2010.
[15] Mittal Amit, Agrawal Ayushi, Chouksey Ayushi, Shriwas Rachna, Agrawal Saloni. A Comparative Study
of Chatbots and Humans. International Journal of Advanced Research in Computer and Communication
Engineering, 2016, 5(3): 1055-1057.
[16] Cheng Xusen, Bao Ying, Zarifis Alex, Gong Wangkun, Mou Jian. Exploring consumers’ response to text-
based chatbots in e-commerce: the moderating role of task complexity and chatbot disclosure. Emerald
Insight Internet Research, 2021, 32(2).
Highlights in Science, Engineering and Technology CMLAI 2023
Volume 39 (2023)
346
[17] Cui Lei, Huang Shaohan, Wei Furu, Tan Chuangqi, Duan Chaoqun, Zhou Ming. SuperAgent: A Customer
Service Chatbot for E-commerce Websites. Association for Computational Linguistics Proceedings of the
55th Annual Meeting of the Association for Computational Linguistics-System Demonstrations, 2017, 97-
102.
[18] Gunawan Dennis, Putri Farica Perdana, Meidia Mira. Bershca: bringing chatbot into hotel industry in
Indonesia. UAD TELKOMNIKA, 18(2): 839-845.
[19] Stoilova Elitza. AI chatbots as a customer service and support tool. ROBONOMICS: The Journal of the
Automated Economy, 2021.
[20] Bates Mary. Health Care Chatbots Are Here to Help. IEEE Pulse, 2019, 10(3): 12-14.
[21] Hsu I-Ching, Yu Jiun-De. A medical Chatbot using machine learning and natural language understanding.
SpringerLink Multimedia Tools and Applications, 2022, 81: 23777-23799.
[22] Winkler Rainer, Matthias Söllner. Unleashing the Potential of Chatbots in Education: A State-Of-The-Art
Analysis. Academy of Management Annual Meeting, 2018.
[23] Jovanović Mladan, Baez Marcos, Casati Fabio. Chatbots as Conversational Healthcare Services. IEEE
Internet Computing, 2021, 25(3): 44-51.
[24] Nuruzzaman Mohammad, Hussain Omar Khadeer. A Survey on Chatbot Implementation in Customer
Service Industry through Deep Neural Networks. IEEE 15th International Conference on e-Business
Engineering, 2018, 54-61.
[25] D’Silva Godson, Jani Megh, Jadhav Vipul, Bhoir Amit, Amin Prithvi. Career Counselling Chatbot Using
Cognitive Science and Artificial Intelligence. SpringerLink Advanced Computing Technologies and
Applications, 2020, 1-9.
[26] Cont Miruna, Ciupe Aurelia, Orza Bogdan, Cohut Irina, Nitu Georgiana. Career Counseling Chatbot using
Microsoft Bot Frameworks. IEEE Global Engineering Education Conference, 2022, 1387-1392.
[27] Ranoliya R. Ranoliya, Raghuwanshi Nidhi, Singh Sanjay. Chatbot for university relate FAQs. IEEE 2017
International Conference on Advances in Computing, Communication and Informatics, 2017, 1525-1530.
[28] Qaffas Alaa A. Improvement o Chatbots Semantics Using Wit.ai and Word Sequence Kernel: Education
Chatbot as a Case Study. Modern Education and Computer Science, 2019, 3: 16-22.
[29] Villanueva Dayana Priscilla Peve, Aguilar-Alonso Igor. A Chatbot as a Support for Educational
Institutions. IEEE 2021 62nd International Scientific Conference on Information Technology and
Management Science of Riga Technical University. 2021, 1-6.
[30] Ho Chan Chun, Lee Ho Lam, Lo Wing Kwan, Lui Kowk Fai Andrew. Developing a Chatbot for College
Student Programme Advisement. IEEE 2018 International Symposium on Educational Technology, 2018,
52-56.
[31] Molnár György, Szüts Zoltán. The Role of Chatbots in Formal Education. IEEE 16th International
Symposium on Intelligent and Informatics, 2018, 197-202.
