1. INTRODUCTION:

Artificial Intelligence is a method to make a computer, or a robot, or a product to reflect how smart human think. AI is a learning of how human brain thinks and act, learn, decide and works. It attempts to solve problems and it outputs the intelligent software systems.

In computer Artificial intelligence sometimes is also referred as machine intelligence where it is an intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans. The term Artificial intelligence is often used to define machines that mimic cognitive functions that humans associate with the human mind, such as learning and problem solving. The objective of AI is to improve computer functions which are related to human knowledge, for example, reasoning, learning, and problem-solving. Artificial intelligence can be classified into three different types of systems: analytical, human-inspired and humanized artificial intelligence. Analytical AI has only characteristics consistent with cognitive intelligence generating cognitive representation of the world and using learning based on past experience to inform future decisions. Human-inspired AI has elements from cognitive and emotional intelligence, understanding human emotions in addition to cognitive elements and considering them in their decision making.

1.1 Working of Artificial intelligence:

Large amount of data is first combined with fast, iterative processing and smart algorithms, which allows the system to learn from the patters in the data. AI is a vast subject and its field of study includes many theories, methods, technology, and it also has major subfields under it. They are:

· Machine Learning: Machine Learning is the learning in which machine can learn by its own from the examples and the experience. The program for this machine need not be specific. The machine tends to change or correct its algorithm from the examples and experiences.

· Artificial Intelligence and Machine Learning are the two most commonly misinterpreted words. They are not the same thing, but the understanding that they are, leads to some confusion. Both these terms arise repeatedly when the topic is Big Data or Data Analytics, or something related to these subjects which is making its rounds around the world.

· Neural Networks: Artificial Neural Networks were inspired by the biological network, i.e. the animal brain. Artificial Neural Networks are one of the most important tools in machine learning to find patterns in the data, which are far too complex for a human to figure out and teach the machine to recognize.

· Deep Learning: In Deep Learning a large amount of data is analysed and the algorithm would perform the task repeatedly, each time twisting/editing the algorithm a little to improve the outcome a little for the better.

· Cognitive Computing: The ultimate goal of cognitive computing is to imitate human thought process in a computer model. Using self-learning algorithms, pattern recognition by neural networks, and natural language processing the computer can mimic the human’s way of thinking. Cognitive computing systems are used to find solutions in complex situations where there are ambiguous and uncertain issues. Here computerized models are deployed to simulate the human cognition process.

· Computer Vision: Computer vision works on allowing computer to see, recognize and process images the same way as the human vision does, and then provides an appropriate output. Computer vision is very closely related with artificial intelligence.

· Natural Language Processing: Natural language processing means communicating with the machines using natural language like English. Machines and algorithms in the workplace are expected to create 133 million new roles, but cause 75 million jobs to be displaced by 2022 according to a new report from the World Economic Forum. This means that the growth of artificial intelligence could create 58 million net new jobs in the next few years.

1.2 Artificial intelligence in banking sector

Artificial Intelligence AI is fast evolving as the go-to technology for companies across the world to personalise experience for individuals. The technology itself is getting better and smarter day by day, allowing more and newer industries to adopt the AI for various applications. Banking sector is becoming one of the first adopters of AI. And just like other segments, banks are exploring and implementing the technology in various ways. The rudimentary applications AI include bring smarter chat-bots for customer service, personalising services for individuals, and even placing an AI robot for self-service at banks. Beyond these basic applications, banks can implement the technology for bringing in more efficiency to their back-office and even reduce fraud and security risks. Unsurprisingly, research firms are bullish on the potential of AI in banking. According to Fintech India report by PwC in 2017, the global spending in AI applications touched $5.1 billion, up from $4 billion in 2015. There is a keen interest in the Indian banking sector as well.

1.3 lists of some common uses of AI in banks:

· Fraud Detection: Anomaly detection can be used to increase the accuracy of credit card fraud detection and anti-money laundering.

· Customer Support and Helpdesk: Humanoid Chatbot interfaces can be used to increase efficiency and reduce cost for customer interactions.

· Risk Management: Tailored products can be offered to clients by looking at historical data, doing risk analysis, and eliminating human errors from hand-crafted models.

· Security: Suspicious behaviour, logs analysis, and spurious emails can be tracked down to prevent and possibly predict security breaches.

· Digitization and automation in back-office processing: Capturing documents data using OCR and then using machine learning/AI to generate insights from the text data can greatly cut down back-office processing times.

· Wealth management for masses: Personalized portfolios can be managed by Bot Advisors for clients by taking into account lifestyle, appetite for risk, expected returns on investment, etc.

· ATMs: Image/face recognition using real-time camera images and advanced AI techniques such as deep learning can be used at ATMs to detect and prevent fraud.

1.4 AI technology in Five leading commercial banks in India:

i. SBI:

SBI, which is India’s largest public-sector bank with 420 million customers and is embarking on its AI journey from the point of view of both employees and customers. SBI is currently using an AI-based solution developed by Chapdex. From a customer chatbot perspective, SBI has launched SIA, an AI-powered chat assistant that addresses customer enquiries instantly and helps them with everyday banking tasks just like a bank representative told Emerj, the company’s own press release for said SAI launch appears to be unavailable due to a website issue, as of Dec 27, 2017. SIA was developed by Payjo, a start-up based in Silicon Valley and Bengaluru. According to Payjo, since its launch, the chatbot has responded to millions of queries from thousands of customers. SIA is setup to handle nearly 10,000 enquiries per second or 864 million in a day. That is nearly 25% of the queries processed by Google every day. Currently, SAI can address enquiries on banking products and services. It is trained with a large set of past customer questions and is said to aptly handle frequently asked questions.

ii. HDFC

HDFC Bank has developed an AI-based chatbot, “Eva”, built by Bengaluru-based Sense forth AI Research. Since its launch in March this year, Eva which stands Electronic Virtual Assistant has addressed over 2.7 million customer queries, interacted with over 530,000 unique users, and held 1.2 million conversations. Eva can assimilate knowledge from thousands of sources and provide simple answers in less than 0.4 seconds Eva has answered more than 100,000 queries from thousands of customers from 17 countries across the globe. With the launch of Eva, the bank’s customers can get information on its products and services instantaneously. It removes the need to search, browse or call. Eva also becomes smarter as it learns through its customer interactions. Currently they have introduced IRA 2.0 that will interact with customers, answer bank-related queries, frequently asked questions, and guide them inside the branch with voice-based navigation.

iii. ICICI:

ICICI Bank, India’s second-largest private sector bank has deployed software robotics in over 200 business processes across various functions of the company. The bank has created the software robotics platform mostly in-house, leveraging AI features such as facial and voice recognition, natural language processing, machine learning and bots among others. I believe that the implementation of software robotics will herald a transformational change in the Indian banking industry. In February this year, ICICI Bank launched its AI-based chatbot, named iPal. Since its launch, the chatbot has interacted with 3.1 million customers, answering about 6 million queries, with a 90 percent accuracy rate.

iv. Axis bank

It is the India’s third-largest private sector bank, launched an innovation lab called Thought Factory last year to accelerate the development of innovative AI technology solutions for the banking sector. The innovation hub located in Bengaluru, has an in-house innovation. Recently, Axis Bank launched an AI & NLP (Natural Language Processing) enabled app, Conversational Banking, to help consumers with financial and non-financial transactions, answer and get in touch with the bank for loan other products. Currently, robotic process automation (RPA) is complete for most processes, including account maintenance and servicing, loan disbursements, bulk transaction processes and ATM support.

v. HSBC:

HSBC is one of the world’s largest banking and financial services organisations, serving around 38 million customers across 67 countries through its 3,900 offices worldwide. HSBC Technology is a function within the Bank that builds and maintains its information technology systems. Technology teams draw on insights from HSBC’s vast global network and combine them with cutting edge financial technology to deliver an unrivalled global banking experience to its customers. HSBC is currently using AI based robots to help in spot money laundering, fraud & also predict how customer might redeem their credit card points.

2. LITERATURE REVIEW:

Amer Awad Alzaidi (2018), the author explores the impact of Artificial intelligence on performance of the banking industry in middle east. This paper aim was to analyse the application of AI in banking sector in middle east and this study presents a comprehensive review of the application of AI techniques in banking sector improving overall performances of the system and banking network. Shivkumar Goel and Nihaal Mehta(2017), the author discuss how AI is used in financial sector and what are the benefit that AI offers to fintech and different ways in which it can improve a financial institute. The key use of AI in FinTech will be augmented decision making. It will allow analysts to make complex decisions with the help of machines which offer both pre and post decision making support, generated by analysing historical data and emerging trends. Bergstrom Stacey, Svenningsson P and Thoresson A (2018), The author explores the attitude that customer have towards AI in customer services, as a substitute to local brick and mortar offices within the Swedish bank industry, as well as uncover any significant factors that could influence these attitudes. Dr. Simran Jewandah (2018) this paper focus on the how AI is changing the banking sector. And study the areas where the AI is being used by the banks and the application of AI in banking sector and use in leading commercial banks in India SBI, HDFC, ICICI, AXIS. T.J.M. Bench-Capon, Paul E. Dunne (2017) this paper focus on the Arguments in AI and to place these contributions in the context of the historical foundations of argumentation in AI and subsequently, to discuss a number of themes that have emerged in recent years resulting in a significant broadening of the areas in which argumentation based methods are used.

2.1 Research Gap:

Majority of research have covered the impact of Artificial intelligence. Very few studies have been carried out on commercial banking sector in India with reference to cost benefit analysis.

3. PROBLEM STATEMENT:

3.1 Aim of the study: To study the implementation of AI in selected commercial banks in India

3.2 Objective:

a) To study the existing AI enabled services available in the banking sector

b) To study the current services in the chosen banks in which AI has been implemented

c) To compare and analyse the AI is used by the leading commercial banks India and across the world

d) To analyse the profitability and feasibility of adopting the AI for the chosen services towards the chosen banks

4. DATA BASE & METHODOLOGIES:

4.1. Data Base:

The data is collected from both secondary and primary basis, for the secondary data various journals and literature review and publications have collected. In this research targeted only on the top 5 leading commercial banks i.e. SBI, ICICI, Axis, HDFC, HSBC of North Bangalore. Based on customer base following banks are selected, SBI has 42 crore, Axis bank has 10 million and ICICI has 5 million, HDFC has 43 million and HSBC has 33 million. Stratified Random sampling has been analysed to collect the responses of the customer, qualitative method has been analysed and based on primary data of the five banks customer response has been taken for the analysis.

4.2 Methodologies:

This study has qualitative method and it is descriptive in nature, for the first objective to study the existing AI enabled services available in the banking sector secondary data was collected by literature review and for the second objective, to study the current services in the chosen banks in which AI has been implemented data collected from Literature Review, Books, Journals, Websites. For the third objective, to compare and analyse the AI is used by the leading commercial banks India and across the world for this Primary data using questionnaire Statistical tools: Correlation, reliability test, Cronbach Alpha, hypothesis, Z – test software and Excel. For the fourth objective, to analyse the profitability and feasibility of adopting the AI for the chosen services towards the chosen banks for this Cost benefit Analysis was analysed the methodologies used for this study is stratified random sampling of 300 sample size, this study targeted on the customer of selected 5 commercial bank in Bangalore. Based on each 11 variable the questionnaire was conducted for survey. From the survey data Reliability test, Factor analysis, Co-relation Descriptive statistics, Frequency test, Hypothesis Z – Test was analysed.

5. RESULT AND DISCUSSION:

Reliability test

Case Processing Summary

		N	%
Cases	Valid	31	100.0
	Excluded^a	0	.0
	Total	31	100.0

a. Listwise deletion based on all variables in the procedure.

Reliability Statistics

Cronbach's Alpha	N of Items
.851	33

5.1: Interpretation:

Cronbach’s α was used to measure the questionnaire’s consistency. The overall coefficient was found to be 0.851 of 33 items which exceeds the minimal recommendations. Therefore, the viability and validity of the instrument is deemed to be sufficient.

5.2 Factor analysis:

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy		0.543
Bartlett's Test of Sphericity	Approx. Chi-Square	787.090
	df	528
	Sig.	0.000

Based on correlations

5.2. Interpretation:

By the data of the customer response in the Questionnaire survey the factor analysis Kaiser- Meyer-Olkin Measure of Sampling Adequacy value is 0.543 which is positive results and Bartlett's Test of Sphericity Approx. Chi-Square value is 787.090 and degree of Freedom value is 528.

5.3 Co-Relation between variables

Table 2 Correlations between market size and capital

		market size	capital
market size	Pearson Correlation	1	-.010
	Sig. (2-tailed)		.870
	N	300	300
capital	Pearson Correlation	-.010	1
	Sig. (2-tailed)	.870
	N	300	300

Interpretation:

From the above table market size and capital are negatively low correlated where the spearman’s value of capital is -.010 and the spearman’s correlation value is 1 which reveals that these two factors are positively correlated as it falls between -1.000 and +1.000 as the value is respondent think AI will increase the market size of the bank and capital investment will neutral.

Table 3 Correlations between capital and growth

		capital	growth
capital	Pearson Correlation	1	0.006
	Sig. (2-tailed)		0.916
	N	300	300
growth	Pearson Correlation	.006	1
	Sig. (2-tailed)	.916
	N	300	300

Table 4 Correlations between few services and growth

		few services	growth
few services	Pearson Correlation	1	-0.043
	Sig. (2-tailed)		0.460
	N	300	300
growth	Pearson Correlation	-.043	1
	Sig. (2-tailed)	.460
	N	300	300

Interpretation:

In this table the privacy and fraud are positively low correlated because the Pearson correlation value is .078 where it comes between 0.01 to 0.35. the spearman’s correlation value is 1 which reveals that these two factors are positively correlated as it falls between -1.000 and +1.000 as the value is positive respondent think AI will have more privacy system for customer data and less fraud.

Interpretation:

In this table of correlation between Growth and capital the spearman’s correlation value is .006 where it is low degree of positively correlated because it comes between 0.01 to 0.35 The spearman’s correlation value is 1 which reveals that these two factors are positively correlated as it falls between -1.000 and +1.000 as the value is positive respondent think AI will increase the growth of banking and capital investment will neutral.

Table 5 Correlations between privacy and fraud

		privacy	fraud
privacy	Pearson Correlation	1	0.078
	Sig. (2-tailed)		0.176
	N	300	300
fraud	Pearson Correlation	.078	1
	Sig. (2-tailed)	.176
	N	300	300

Interpretation:

In this table of correlation between Few services and growth the spearman’s correlation value is -.043 where it is low degree of negatively correlated because it comes between 0.01 to 0.35 The spearman’s correlation value is 1 which reveals that these two factors are positively correlated as it falls between -1.000 and +1.000 as the value is respondent think few new AI services should not be made mandatory for customer benefit.

5.5 Descriptive Statistics:

Table 8

	N	Minimum	Maximum	Mean	Std. Deviation
market size	300	1	5	1.93	0.912
quick	300	1	5	1.69	0.633
satisfy	300	1	5	2.03	0.712
capital	300	2.0	5.0	3.787	0.8739
revenue	300	1	4	2.25	0.656
growth	300	1	5	2.86	1.143
easy	300	1	4	1.97	0.792
reduction	300	1	5	1.71	0.579
machine	300	1	4	1.82	0.693
age	300	1	5	1.64	0.678
need	300	1	5	2.32	0.621
experience	300	1	5	1.66	0.730
easier	300	1	5	2.05	0.697
network	300	1	4	2.17	0.685
communication	300	1	5	1.79	0.745
privacy	300	1	5	2.90	1.024
few services	300	1	5	2.68	0.872
fraud	300	1	5	3.44	0.981
innovative	300	1	4	1.72	0.628
long run	300	1	3	2.33	0.633
time	300	1	4	1.66	0.616
chat box	300	1	5	1.57	0.605
use of	300	1	5	2.06	0.729
less risk	300	1	5	2.77	0.861
decision	300	1	5	1.91	0.716
digitalize	300	1	5	1.63	0.703
working condition	300	1	5	2.48	0.667
unemployemet	300	1	3	1.67	0.563
man power	300	1	5	1.62	0.635
competition	300	1	5	2.25	0.940
aware	300	1	5	1.61	0.626
knowledge of use	300	1	3	1.57	0.559
positively	300	1	3	1.53	0.520
Valid N (listwise)	300

Interpretation:

In this table we can see the mean and standard deviation as well as sample size for each of the item on my scale and this is same as if we would run on frequency or descriptive statistics produce some mean and standard deviation.

5.5 Cost benefit analysis:

A cost benefit analysis also known as a benefit cost analysis is a process by which organizations can analyse decisions, systems or projects or determine a value for intangibles. The model is built by identifying the benefits of an action as well as the associated costs and subtracting the costs from benefits. A cost-benefit analysis can be a useful tool for decision-making, but the accuracy of a cost-benefit analysis is limited by the thoroughness of recognizing likely costs and benefits. If a business fails to recognize potential costs and benefits, it can cause poor results that lead to sub-optimal decisions.

Table 9 Benefit cost

BANK	BENEFIT COST
SBI	2,78,083
ICICI	77,913
AXIS	68,116
HDFC	1,16,597
HSBC	96,461

Table 10 Benefit cost ratio

BANK	BENEFIT COST RATIO
SBI	7.95182
ICICI	2.22943
AXIS	1.94909
HDFC	3.3341096
HSBC	2.58317

Interpretation:

A cost-benefit analysis is a process business use to analyse decisions. The business or analyst sums the benefits of a situation or action and then subtracts the costs associated with taking that action. To identify the benefit cost ratio of AI technology in banking sector this calculation has been analysed. According to fintech India report by PWC in 2017, the global is spending in AI application touched $5.1 billion up from $4 billion in 2018. By this analysis estimation of cost can be invested for AI application banking sector. SBI can invest ₹ 7.95182 and ICICI can invest ₹2.22943 and AXIS can invest ₹1.94909 and HDFC can invest ₹ 3.3341096 and ₹ 2.58317 in India.

6. CONCLUSION:

A digitalization is certainly taking place across all segments of industry especially banking. Artificial intelligence is the field of science that deals with rivalling the capabilities of modern computer systems to resolve issues using human-like complex capabilities of reasoning, learning and self-correction. Currently artificial intelligence is used in detecting mismatching in transactions, providing personalised recommendations for the customers and developing solution for eliminating human errors. The traditional banking has evolved and more and more banks are adopting new technologies like AI, Cloud, block chain to cut down their operating expenses and improve efficiency. Improvement and development in the AI industry will increase productivity at a reduced cost. The banks are increasingly looking at emerging technologies such as blockchain and analytics in creating an active defense mechanism against cybercrimes. Artificial Intelligence could be the key to transforming many of these crucial customers facing processes and retaining the competitive edge. The bank should focus on making aware of their customer about the AI technology and benefit of the technology. This study uses stratified random sampling technique for the data collection. As result of the survey majority of the customer are accepting the new AI technology in banking sector but lacking is here most of the customer are not aware of the new technologies in the market and usage of the technologies. It necessary for the bank to focus on make aware of knowledge of the technologies and usage of it.

7. LIMITATION:

The study is restricted to 5 selected commercial banks in India.

8. FUTURE RESEARCH:

This study is restricted to only Top 5 Indian banks in north Bangalore based on customer base. This research focused only on selected banking customer of Bangalore by using stratified random sampling. Require further research to identify the other factors that promote the adoption of AI in banking sector in city. It is necessary to search variables that increase our ability to better understand the actual use and predictability of intended use.

9. REFERENCES:

1. M. Bhuvana, P. G. Thirumagal and S .Vasantha, Big Data Analytics - A Leveraging Technology for Indian Commercial Banks, Indian Journal of Science and Technology, Vol 9 (32), August 2016

2. Hoikkala, H. & Magnusson, N. (2017). Swedish banks embrace artificial intelligence as a cure to closures. The Independent. Retrieved 26 Februari, 2018

3. Dusko Knezevic,(2017) ”Impact of Artificial intelligence in Changing the Financial Sector and Other Industries” Montenegrin Journal of Economics Vol. 14, No. 1 (2018), 109-120

4. Lata Varghese, Manish Tomer, Fletcher McCraw, (2017), “ A study on Artificial intelligence technology” International Journal of Engineering & Technology, 7 (2.7) (2017) 418-421

5. Eisazadeh, S., Shaeri, Z. and Ali, B. (2012) ‘An analysis of bank efficiency in the middle east and north africa’, The International Journal of Banking and Finance, 9(4), pp. 28–47.

6. Dabholkar, P. A. (2016). Consumer evaluations in new technology-based self-service options: an investigation of alternative model of service quality, International Journal of Research in Marketing, Vol. 13 (1), pp. 29-51.

7. Dabholkar, P. A., & Bagozzi, R. P. (2017). An attitudinal model of technology-based selfservice: moderating effects of consumer traits and situational factors. Journal of the Academy of Marketing Science, Vol. 30 (3), pp. 184-201

8. Hanafizadeh, P., Keating, B. W., & Khedmatgozar, H. R. (2014). A systematic review of AI technology banking adoption. Telematics and Informatics, Vol. 31 (3), pp. 492-510

9. Mata, J. et al. (2018) ‘Artificial intelligence (AI) methods in optical networks: A comprehensive survey’, Optical Switching and Networking, 28, pp. 43–57

10. Manning, J. (2018) How AI Is Disrupting the Banking Industry, International Banker. Available at: https://internationalbanker.com/banking/how-ai-is-disrupting-thebanking-industry

11. Mannino, A. et al. (2015) ‘Artificial Intelligence: Opportunities and Risks Policy paper’, Effective Altruism Foundation. Available at: www.foundational-research.org.

12. AI application in 5 years available in applications-in-banking-to-look-out-for-in-next-5-years

13. AI in future banking sector available in https:// complyadvantage.com/artificialintelligence-will-drive-the-future-of-fintech

Received on 04.09.2019 Modified on 20.10.2019

Asian Journal of Management. 2019; 10(4):377-384.

DOI: 10.5958/2321-5763.2019.00057.X

0.831661093	0.763701226	1.04948718	0.962531	1.307146	0.760134
z-Test: Two Sample for Means

	Market size	capital
Mean	1.933333333	3.78666667
Known Variance	0.831661	0.763701
Observations	300	300
Hypothesized Mean Difference	0
z	-25.41467396
P(Z<=z) one-tail	0
z Critical one-tail	1.644853627
P(Z<=z) two-tail	0
z Critical two-tail	1.959963985

	privacy	fraud
Mean	2.896666667	3.43666667
Known Variance	1.04948	0.962531
Observations	300	300
Hypothesized Mean Difference	0
z	-6.593852253
P(Z<=z) one-tail	2.142785949
z Critical one-tail	1.644853627
P(Z<=z) two-tail	4.285571897
z Critical two-tail	1.959963985

	growth	few services
Mean	2.856666667	2.68
Known Variance	0.962531	0.760134
Observations	300	300
Hypothesized Mean Difference	0
z	2.331389452
P(Z<=z) one-tail	0.009866417
z Critical one-tail	1.644853627
P(Z<=z) two-tail	0.019732833
z Critical two-tail	1.959963985