A Study on the Implementation and the Impact of Artificial Intelligence in Banking Processes

 

Rashmi R.¹, Nirmal Raj VK²

¹Assistant Professor, Faculty of Management and Commerce, MS Ramaiah University of Applied Sciences, Bengaluru, Karnataka, India–560054.

²MBA Student, Faculty of Management and Commerce, MS Ramaiah University of Applied Sciences, Bengaluru, Karnataka, India–560054.

 

ABSTRACT:

Artificial intelligence is a broad branch of computer science. When we think about Artificial intelligence, we need to think in the context of human beings the reason being human beings are known to be the smartest creature in the world. So, when technology like Artificial intelligence is brought for digitalisation of banks then competing with fin-tech players becomes easy. The main reason for me to do artificial intelligence as a topic was due to the fact that A.I and machines are the future. The main focus I kept on A.I on banking processes, reason being the number of scams happening in India and around the world, also lack of technology being the main culprit for fraudulent activities and scams. To analyse the impact of AI on the Banks’ Performance, a questionnaire is prepared with 34 variables which included different banking processes. The respondents included a combination a employees and customers. Through the survey most of the variables were showing positive results towards adopting Artificial Intelligence in banking sector. According to the survey analysis the reliability coefficient of Cronbach’s alpha is 0.972, which indicates high level of internal consistency of the scale and the (KMO) Kaiser-Meyer-Olkin and Bartlett’s test revealed the measure of sampling adequacy is 0.960, it is found that component analysis is useful and significant. As per the regression model the independent variable that is understanding customer behaviour has least impact on the dependent variable that is the impact of Artificial Intelligence on Banks performance. And the independent variable that is customer satisfaction- help desk has great impact on the dependent variable.The Significant values of the coefficients that is the independent variables - Customer satisfaction, Eliminating Human errors, Risk management and automate compliance are significant as the p-value is less than 0.05 (p<0.05).

 

 

 

 

INTRODUCTION:

Artificial intelligence is a broad branch of computer science. When we think about Artificial intelligence, we need to think in the context of human beings the reason being human beings are known to be the smartest creature in the world.

 

The way humans communicate through language, Artificial intelligence uses speech recognition for doing the same and it is based on statistics, hence known as statistical learning.

 

The way humans write, head text in a language, in the field of Artificial intelligence it is known as NLP (Natural language processing).

 

Humans can also see with their eyes and process what they see and the same process in A.I is known as computer vision. Image processing is equally important.

Digital disruption has redefined how the industry should work and has changed the way how business usually functions.

 

Banking industry has changed a lot too and has becomes more customer centric.

 

As customers are becoming more tech savvy, they expect the bank to provide them the seamless experience as well. The seamless experience like mobile banking, e- banking and real time money transfers have made customers life easy, but have cost banking sector huge investments.

 

Robotic automation process enables 80% of repetitive work processes, so that the knowledge workers can dedicate their work and money towards value added services which requires high level of human intervention.

 

To conclude, Artificial intelligence is not a futuristic concept for banks anymore it is a necessity which should be implemented for the present and future.

 

Literature Review:

The era of artificial intelligence in Swedish banking, the artificial intelligence is replacing jobs and traditional services and the prediction is that in coming years they are going to become the biggest market trend.Four major Swedish banks after implementing A.I have experienced severe decline in customer satisfaction, one of the main reasons is decline of local branches. (Stacey Isabel et al., 2018)

 

An AI- based, multi- stage detection system of banking botnets, the primary drivers of finance related cybercrime are banking Trojans and botnets, compared to the baseline models deep learning on detections are more successful.Cybercriminals from past experience are learning and adapting quickly and also taking help of highly sophisticated technologies.Some of the financial organizations have started to take cyber kill chain (CKC). For supporting taxonomies to investigate tactics and to perform day-to-day tasks APT. (Zbiqiang et al., 2019)

 

Artificial Intelligence powered Banking Chatbot focus on the use of chatbots in Banking. Chatbot is an artificial robot created and designed in order to conversate with customers. Chatbots are more likely a virtual assistant. In customer service departments in banks humans cannot always be available but bots can be available 24/7. This study has extracted a contrast between humans and chatbots. This paper’s data collected for the analysis is the FAQs on the bank websites and technologies used in banking for customer care. This paper has also discussed the comparison of seven ML algorithm classification which is used for getting input class to chatbot. (Kumar Satheesh et al., 2018)

 

Objective of the study:

To analysis and evaluate the impact of Artificial Intelligence on Bank’s performance through Survey method

 

METHODOLOGY:

In this study the implementation of Artificial Intelligence in banking processes is studied thoroughly. Individually, the banking processes where in artificial intelligence is implemented are examined. Contrast between human and Artificial Intelligence in different banking processes with respect to the pros and cons are studied and reviewed. In order to analyze the impact of artificial intelligence on Bank’s performance, questionnaire survey method is used and the respondents are the employees of banks (Back-end employees, Vice – President, Front – desk agents, etc.) and the customers of banks.

 

The following steps are involved:

·       To analyse the inputs of questionnaire survey using Correlation, Descriptive Statistics, Factor analysis and Reliability tests on SPSS.

·       To evaluate the impact of Artificial Intelligence on Bank’s performance using Regression Model (Linear and Auto Linear Regression Model).

 

RESULTS AND DISCUSSIONS:

I have created a questionnaire with 30 variables including the demographic profile of the respondents. The main variables that is a set of 21 are used to analyse the if there is (significant) positive or (insignificant) negative impact on Bank’s performance due to the implementation of Artificial Intelligence.

 

The respondents were the employees of the Banks and the customers of banks. The sample size of my survey, that is n = 200.

 

Kaiser – Meyer – Olkin and Bartlett’s Test:

Table No. 1: Result of KMO and Bartlett’s Test

KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.		.960
Bartlett's Test of Sphericity	Approx. Chi-Square	4593.688
	df	276
	Sig.	.000

 

 

 

KMO and Bartlett's Test:

In the Table No. 1, we see KMO (Kaiser-Meyer- Olkin) measure of sampling adequacy is 0.960 which shows that components analysis is useful and survey is significant. The value of KMO should be greater than 0.6 for it to be useful. This measure varies between 0 and 1, and values closer to 1 are better. The survey analysis value is 0.960. A value of 0.6 is a suggested minimum.

Table No. 2: Case Processing Summary of Reliability Analysis

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

 

ANOVA with Friedman’s Test:

 

Table No. 3: Result of ANOVA with Friedman’s Chi-Square Test

ANOVA with Friedman's Test
		Sum of Squares	df	Mean Square	Friedman's Chi-Square	Sig
Between People		2823.591	199	14.189
Within People	Between Items	993.812a	23	43.209	1617.536	.000
	Residual	1832.422	4577	.400
	Total	2826.234	4600	.614
Total		5649.825	4799	1.177

Grand Mean = 1.8443228

a. Kendall's coefficient of concordance W = .176.

 

Reliability Analysis:

Table No. 4: Result of Reliability Analysis

 

Reliability Analysis:

The Reliability coefficient of Cronbach’s alpha normally ranges between 0 and 1.

 

Closer the reliability coefficient is to 1.00, the greater is the internal consistency of the items (variables) in the scale.

 

As per Table No. 4, reliability analysis output derived, the Cronbach’s alpha is 0.972, which indicates a high level of internal consistency for our scale with the specific sample.

 

Cronbach’s alpha coefficient increases either as the number of items (variables) increases, or as the averages inter – item correlations increase (that is when the number of items is held constant).

 

Developing Regression Model:

Assumptions considered for the Regression model - The following are the assumptions considered in regression modelling:

·       The Dependent variable used for the regression modelling is Impact of Artificial Intelligence on Banks’ performance.

·       The Independent variables for the regression modelling are Customer satisfaction, Understand Customer Behaviour, Risk management, Personalised products, Performing Risk analysis and Decision making, Eliminating Human Errors, Asset management and wealth management, Enhance Portfolio management, Automate Compliance, Reduction in cost of services, Ease in access and ease in services, Fraud detection and suspicious behaviour, Anti–Money Laundering services, Security, Tracking spurious emails and Log analysis, Predict security breaches, Digitization and automation in back-office processing and Enhanced performance of ATM’s.

 

Table No. 5: Result of Auto Linear Regression Model Summary

 

 

Interpretation of Auto Linear Regression Model Summary:

As per the Table No. 5, the smaller the information criterion, the better is the model fit. The accuracy of the model is greater than 60%, it is found that the model fits better with the accuracy of 71.9%.

 

Figure No. 1: Output of Predictor Importance

 

Interpretation of Predictor Importance:

As per the Figure No. 1, as per the survey, the Independent variable that is Reduction in cost of services and Understanding Customer Behaviour is least important that means it has least impact on the Dependent variable that is the Impact of Artificial Intelligence on Banks’ performance. And the Independent variable that is Customer Satisfaction – help desk is most important that means it has great impact on the Dependent variable that is the Impact of Artificial Intelligence on Banks’ performance.

 

Figure No. 2: Normal Distribution Plot

 

Table No. 6: Effects of the Dependent variables on the Independent variable

 

 

Figure No. 3: Diagrammatic representation of Effects of the Dependent variables on the Independent variable

 

Interpretation of the Effects:

In the Table No. 6 and Figure No.3, as per the survey, the Independent variable that is Reduction in cost of services and Understanding Customer Behaviour has least impact on the Dependent variable that is the Impact of Artificial Intelligence on Banks’ performance. And the Independent variable that is Customer Satisfaction–help desk has great impact on the Dependent variable that is the Impact of Artificial Intelligence on Banks’ performance. The significance value is 0.000 that means and is found that it is statistically significant.

 

Table No. 7: Coefficients of Independent and Dependent variables

 

Interpretation of Coefficients of Auto Linear Regression Model:

In our Auto Linear Regression analysis, the test conducted tests the null hypothesis that the coefficient is o (zero), the t-test depicts that both the variable and the intercept are highly significant (p<0.0001) and thus we might say the they are different from zero. The Significant values of the coefficients that is the independent variables-Customer satisfaction, Eliminating Human errors, Risk management and automate compliance are significant as the p-value is less than 0.05 (p<0.05) it is found that these variables have significant impact on the dependent variable – Impact of Artificial Intelligence on Banks’ performance. whereas independent variable that is understand customer behaviour is not significant that is p = 0.092 which is p>0.05 it is found that these variables have insignificant impact on the dependent variable–Impact of Artificial Intelligence on Banks’ performance.

 

Table No. 8: Model Building Summary

 

Figure No. 4: Diagrammatic representation of Independent variables’ estimated means

 

As per the Figure No. 4 and Table No. 7, if we look at the p-values of the estimated coefficients, we can see and find out that not all the coefficients are statistically significant (p<0.005). this means that only a subset of the predictors (Independent variables) are related to the outcome or the results.

 

Linear Regression Model:

Table No. 9: Result of Linear Regression Model

Model Summary
Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	.842a	.709	.681	.491

 

a. Predictors: (Constant), Enhanced performance of ATM's by use of Image/face recognition using real-time camera images and advanced Artificial Intelligence techniques, Risk Management, Automate compliance, Asset management and Wealth management, Fraud Detection and Suspicious behavior, Understand Customer Behavior, Performing risk analysis and Decision making, Personalized products being offered to customers by looking at historical data, Customer Satisfaction (Customer-support and help desk), Eliminating human errors, Anti - Money Laundering service, Reduction in cost of services, Digitization and automation in back-office processing, Enhance portfolio management, Tracking spurious emails and Log analysis, Security , Predict security breaches, Ease in access and Ease of services

 

ANOVAa
Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	106.696	18	5.928	24.552	.000b
	Residual	43.699	181	.241
	Total	150.395	199

a. Dependent Variable: Banks' Performance

 

b. Predictors: (Constant), Enhanced performance of ATM's by use of Image/face recognition using real-time camera images and advanced Artificial Intelligence techniques, Risk Management, Automate compliance, Asset management and Wealth management, Fraud Detection and Suspicious behavior, Understand Customer Behavior, Performing risk analysis and Decision making, Personalized products being offered to customers by looking at historical data, Customer Satisfaction (Customer-support and help desk), Eliminating human errors, Anti - Money Laundering service, Reduction in cost of services, Digitization and automation in back-office processing, Enhance portfolio management, Tracking spurious emails and Log analysis, Security, Predict security breaches, Ease in access and Ease of services

 

 

Coefficientsa
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	Std. Error	Beta
1	(Constant)	.015	.098		.158	.875
	Customer Satisfaction (Customer-support and help desk)	.374	.070	.397	5.382	.000
	Understand Customer Behavior	.116	.064	.131	1.808	.072
	Risk Management	.138	.063	.145	2.195	.029
	Personalized products being offered to customers by looking at historical data	-.011	.058	-.013	-.193	.847
	Performing risk analysis and Decision making	.008	.060	.010	.135	.893
	Eliminating human errors	.160	.056	.209	2.844	.005
	Asset management and Wealth management	.063	.059	.079	1.061	.290
	Enhance portfolio management	.013	.062	.017	.214	.831
	Automate compliance	.084	.060	.105	1.406	.161
	Reduction in cost of services	.037	.064	.045	.581	.562
	Ease in access and Ease of services	.031	.065	.040	.483	.630
	Fraud Detection and Suspicious behavior	-.060	.063	-.077	-.951	.343
	Anti - Money Laundering service	-.085	.060	-.107	-1.415	.159
	Security	.034	.066	.043	.517	.606
	Tracking spurious emails and Log analysis	.059	.065	.074	.910	.364
	Predict security breaches	-.002	.068	-.003	-.031	.975
	Digitization and automation in back-office processing	-.050	.063	-.064	-.804	.422
	Enhanced performance of ATM's by use of Image/face recognition using real-time camera images and advanced Artificial Intelligence techniques	-.055	.062	-.068	-.882	.379

a. Dependent Variable: Banks' Performance

 

Interpretation of Linear Regression Model:

The coefficient of determination is 0.709; therefore, about 70.9% of the variations in the Independent variables are explained by the dependent variable that is the Impact of Artificial Intelligence on Bank’s performance. The regression equation appears to be very useful for making predictions since the value of R^2 (R square) is close to 1.

 

The higher the R-square value, Adj - R square value and F-statistic value, the better it is (>0.60).

 

CONCLUSION:

Artificial Intelligence impact on banking processes and how it satisfies customers better when compared to traditional banking services. A.I experiences like chatbots, online banking and robo- advisors are able to serve customers better specially the new age millennials do not have the time to be physically available in a bank branch, instead they would sit at their home or work place to do the task online.

 

Banking industry came into A.I late compared to other industries and they came in to drive the banking business further.

 

The initial steps were to understand and study the importance and implementation of A.I in banking industry. Then a proper contrast was set up to compare which among human employees and A.I is better for particular banking processes.

 

According to the survey analysis the reliability coefficient of Cronbach’s alpha is 0.972, which indicates high level of internal consistency of the scale and the (KMO) Kaiser-Meyer-Olkin and Bartlett’s test revealed the measure of sampling adequacy is 0.960, it is found that component analysis is useful and significant.

 

As per the regression model the independent variable that is understanding customer behaviour has least impact on the dependent variable that is the impact of Artificial Intelligence on Banks performance. And the independent variable that is customer satisfaction- help desk has great impact on the dependent variable.

 

The Significant values of the coefficients that is the independent variables-Customer satisfaction, Eliminating Human errors, Risk management and automate compliance are significant as the p-value is less than 0.05 (p<0.05)

 

ACKNOWLEDGEMENT:

It is great pleasure for me to take up this project. I feel honored doing the project entitled - “A Study on the implementation And the Impact of Artificial Intelligence in Banking Processes”.

 

I am grateful to my project guide Mrs. Rashmi R assistant professor of MS Ramaiah University of Applied Sciences and also Dean DR. Srivatsa of MS Ramaiah University of Applied Sciences. I am thankful to them as through the journey of this project I did come across so many new things which has improved knowledge.

 

This project wouldn’t have been completed without their precious help, guidance and worthy experience. Whenever I was in a need or required guidance, they were with me.

 

Although this report has been carried out with utmost care and interest, I am ready to accept respondent and imperfection.

 

REFERENCES:

1.     Bergström Stacey, I., Svenningsson, P., andThoresson, A. (2018). The Era of Artificial Intelligence in Swedish Banking: Exploring Customer Attitudes Towards AI as a Substitute to Brick and Mortar Offices, 86. Retrieved from http://www.diva-portal.org/ smash/record.jsf?pid=diva2%3A1213837&dswid=2098

2.     Ling, L., Gao, Z., Silas, M. A., Lee, I., (2019). An AI-based, Multi-stage detection system of banking botnets. Retrieved from http://arxiv.org/abs/1907.08276

3.     Ukpong, E. G., Udoh, I. I., and Essien, I. T. (2019). Artificial Intelligence: Opportunities, Issues and Applications in Banking, Accounting, and Auditing in Nigeria. Asian Journal of Economics, Business and Accounting, 1–6. https://doi.org/10.9734/ajeba/2019/ v10i130099

4.     Nitescu, d. C., (2018). Managing digitalization in banking: challenges and implications. proceedings of the 12th international management conference: management perspectives in the digital era (imc 2018) (pp. 339–349).

5.     Ivanov, S., and Webster, C. (2017). Adoptıon Of Robots, ArtıfıcıalIntellıgence And ServıceAutomatıon By Travel, Tourısm And HospıtalıtyCompanıes – A Cost-BenefıtAnalysıs. International Scientific Conference “Contemporary Tourism – Traditions And Innovations (pp. 168–177). 168-179. Retrieved from https:// sozluk.gov.tr/?kelime=dijital

6.     Ortiz, J., Marin, A., (2016). Implementation of a banking system security in embedded systems using artificial intelligence. Advances in Natural and Applied Sciences, 10(17), 95–101.

7.     Nataraja N.S, Nagaraja Rao Chilale, Ganesh L. An Examination of Applicability of Logistic Regression Model with Respect to Ready-To-Eat food Products. Asian J. Management; 2017; 8(4):1351-1355.

8.     Sindhu J, Renee Namratha. Impact of Artificial Intelligence in chosen Indian Commercial Bank –A Cost Benefit Analysis. Asian Journal of Management. 2019; 10(4):377-384.

 

Accepted on 26.11.2020           ©AandV Publications All right reserved