Chatbots and young people in emerging economies: Factors affecting user satisfaction

Figures & data

Figure 1. Research model proposed in this study to analyse the use of chatbot among young university students.

Table 1. Items of the questionnaire used in the study

Factor	Item	Description:
Knowledge of use	KU1	Knowing that virtual assistance in my financial institution influences whether I use the service
	KU2	Knowing concepts and skills of using the chatbot helps me use it correctly
	KU3	Knowing how to handle the virtual assistance service helps me solve my issues without having to leave home
	KU4	Getting educated about the chatbot influences my use of it
Attention to natural language	ANL1	The language used by the chatbot is compatible with the service
	ANL2	I feel comfortable interacting with the virtual assistant, as if I were interacting with a human
	ANL3	Robotic language makes it difficult to resolve my banking issue
Access to updated information	AUI1	The information I receive is updated and according to the concern presented
	AUI2	The chatbot has questions about the most common topics so that I do not need to leave the house
	AUI3	I receive a consistent and comprehensive response to all my concerns
Resolution of pending cases	RPC1	It is easy to receive a response when I use the chatbot in my financial institution
	RPC2	I receive clear answers and additional options to the mentioned topic
	RPC3	I do not have to repeat my concerns since the chatbot stores the conversation history, and thus, my question is resolved

Source: the Authors.

Table 2. Sociodemographic questions and complementary information

Age range	Percentage	Level of education	Percentage
18 to 26 years	31.50%	Technical	31.50%
27 to 35 years	45.70%	Technology	38.60%
36 to 40 years	16.50%	Professional	23.60%
41 years or older	6.30%	Postgraduate	6.30%
Occupation	Percentage	Knowledge of what a chatbot/virtual assistant is	Percentage
Employee	76.40%
Student	14.20%
Independent Worker	6.30%
Unemployed	1.60%	No	9.40%
Other, which?	1.60%	Yes	90.60%
Importance of chatbots for financial services	Percentage	Income range *SMMLV: Current Minimum Monthly Legal Salary	Percentage
Very important	3.10%
Important	4.70%	Less than 1 SMMLV	15.70%
Neutral	15.70%	Between 1 and 3 SMMLV	69.30%
Little importance	22.80%	Between 4 and 5 SMMLV	11.80%
Not important	53.50%	More than 5 SMMLV	3.10%
Type of financial products used	Percentage	Level of satisfaction when accessing virtual assistance	Percentage
Credit card	29.90%	Excellent	39.40%
Credits	11%	Good	26.80%
Investment funds	0.80%	Regular	19.70%
Mortgage loan	2.40%	Bad	9.40%
Savings or checking account	55.90%	Very bad	4.70%
Expectations about the virtual assistance service	Percentage	Expectations about the virtual assistance service	Percentage
Resolve the issue	51.20%	Have a similar language	20.50%
Quick service	40.90%
Effectiveness	43.30%	Coherent answer	0.80%

Source: the Authors.

Table 3. Cross factor loadings

	Knowledge of use	Attention to natural language	Access to updated information	Resolution of pending cases
KU1	0.895	0.583	0.633	0.612
KU2	0.704	0.418	0.415	0.371
KU3	0.881	0.586	0.594	0.571
KU4	0.839	0.451	0.453	0.448
ANL1	0.566	0.906	0.701	0.768
ANL2	0.543	0.881	0.651	0.647
AUI1	0.572	0.700	0.946	0.848
AUI2	0.558	0.617	0.876	0.715
AUI3	0.637	0.764	0.939	0.895
RPC1	0.592	0.768	0.873	0.927
RPC2	0.608	0.761	0.864	0.954
RPC3	0.487	0.655	0.722	0.876

Source: The Authors.

Table 4. Convergent validity of the model

Construct	Indicator	Factor loading	VIF	CA	CR	AVE
Knowledge of use	KU1	0.895	2.550	0.852	0.880	0.694
	KU2	0.704	1.447
	KU3	0.881	2.433
	KU4	0.839	2.135
Attention to natural language	ANL1	0.906	1.556	0.748	0.755	0.798
	ANL2	0.881	1.556
Access to updated information	AUI1	0.946	4.339	0.910	0.922	0.848
	AUI2	0.876	2.357
	AUI3	0.939	3.926
Resolution of pending cases	RPC1	0.927	3.656	0.909	0.919	0.846
	RPC2	0.954	4.736
	RPC3	0.876	2.468

Source: The Authors.

Table 5. Fornell-Lacker criterion

	Access to updated information	Attention to natural language	Knowledge of use	Resolution of pending cases
Access to updated information	0.921
Attention to natural language	0.757	0.894
Knowledge of use	0.641	0.621	0.833
Resolution of pending cases	0.895	0.795	0.614	0.920

Source: The Authors.

Table 6. Hypothesis testing

Hypothesis		Path value	T value	p value	Acceptance
Knowledge of use → Access to updated information	H1	0.129	1.719	0.086	Not accepted
Attention to natural language → Access to updated information	H2	0.079	0.771	0.440	Not accepted
Resolution of pending cases → Access to updated information	H3	0.664	9.607	0.000	Accepted
Knowledge of use → Resolution of pending cases	H4	0.614	11.392	0.000	Accepted
Resolution of pending cases → Attention to natural language	H5	0.753	8.061	0.000	Accepted
Knowledge of use → Attention to natural language	H6	0.213	2935	0.003	Accepted

Source: The Authors.

Figure 2. Model of chatbot use among young university students. Note: *R²; **p value. Source: the Authors.

Figure 3. Results of the model produced by the SmartPLS program.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.