Evaluating the Effectiveness of Robo-Advisory in Retail Investments: A Case of Malaysia

4.3. Paired Sample T-test

A paired-samples t-test was conducted to examine whether portfolio returns and portfolio diversification score increased when the group of investors was exposed to robo-advisory services. The results are specified in Table 2.

Table 2. Paired sample T-test.

The results of the paired-samples T-test indicate a statistically significant increase in portfolio returns after the robo-advisory. The thirty-day simulated portfolio returns were surged from 4% to nearly 7% following the intervention of robo-advisory. The difference was also statistically significant at (t = –21.21; p < 0.001). As the magnitude of surges is found to be modest in absolute terms, its economic pertinence must be interpreted within the short simulation horizon and regulated or controlled conditions. Such increments in the returns can be transformed into materially greater cumulative returns. The t-test value is negative because the paired difference (Pre – Post) is negative when post-intervention returns exceed pre-intervention returns. The negative sign therefore indicates improvement, not deterioration. To further assess practical importance, Cohen’s d was calculated to estimate the size of the pre-post change, indicating a large effect size consistent with the statistical significance of the findings. In addition, when taken into account alongside the observed portfolio diversification, the results indicate meaningful optimisation rather than just numerical gain. However, the simulated setting limits direct inference about long-run wealth accumulation, warranting careful interpretation of its economic significance. These findings suggest that post-advisory returns are considerably higher than pre-advisory returns. In general, the results clearly show that a robo-advisory intervention significantly improved investment performance over the simulated 30-day period.

From the results in Table 3, related to the mean diversification score pre- and post-robo-advisory intervention, it can be observed that the pre-experiment mean (0.367, SD: 0.081) was significantly increased to 0.603 (SD: 0.191). This indicates that robo-advisory services significantly increased portfolio diversification scores. The Maximum score was also increased from 0.52 to 0.894. The diversification score was computed by applying a normalised index that captures asset class dispersion and weight balance within the portfolio holdings, where higher values indicate greater diversification. The score, which is closer to zero, mimics concentrated portfolios, whereas values above 0.60 are usually taken to indicate portfolios that are effectively diversified. From a theoretical perspective, this enhancement is consistent with MPT, which posits that diversification reduces unsystematic risk without proportionately decreasing expected returns. The post-intervention indicates that assets were reallocated systematically through a robo-based investment advisory to achieve a more even risk distribution than in participants’ initial portfolios. Considering that retail investors in Malaysia usually exhibit concentration tendencies and home bias, the observed shifts towards greater diversification suggest a return-maximising portfolio optimisation rather than random variation. However, as the score is based on a simulated setting, it indicates structural improvements in asset allocation quality rather than guaranteed long-term risk reduction.

Table 3. Mean diversification score.

4.4. Measurement Model Confirmatory Factor Analysis (CFA)

The reliability, internal consistency, and convergent validity of the study’s measurement model are analysed using CFA approaches, including Cronbach’s Alpha, Composite reliability, and convergent validity, as indicated by (Haji-Othman & Yusuff, 2022) as well. The results of CFA are provided in Table 4.

Table 4. Measurement model.

Results of the CFA indicate valid and reliable measurement model. The factor loadings for all constructs are high, ranging from 0.814 to 0.934, which exceeds the acceptable minimum of 0.60 and suggests that all items represent the latent variable to a high degree. Cronbach’s alpha and composite reliability are also high, above the threshold of 0.70, indicating high internal consistency and reliability. Convergent validity is also supported, with all AVEs exceeding 0.50 and values ranging from 0.733 to 0.835, indicating that each construct accounts for a significant percentage of the variance in its indicators. The measurement model, in general, has high psychometric properties, indicating that the Investor Digital Literacy, Perceived Fairness, Perceived Usefulness, Trust in Algorithm, and Robo-Advisor Adoption and Personalisation are measured reliably and validly.

4.5. Discriminant Validity

The discriminant validity explains the separability and distinctiveness among the constructs of the study and also allows analyse of whether variables do not possess conceptual overlap. It is analyse using HTMT ratio against the threshold of 0.85. The results of discriminant validity are provided in Table 5.

Table 5. Discriminant validity.

The HTMT scores indicate that discriminant validity is generally achieved across all constructs. All HTMT ratios range between 0.235 and 0.585, well below the 0.85 threshold, confirming discriminant validity. For instance, the HTMT value is found between Perceived Usefulness and Robo-Advisory Adoption and Personalisation (0.305) which shows no conceptual overlapping and distinctiveness among variables. Similarly, the HTMT value between Investor Digital Literacy and Perceived Fairness (0.401) is lower than the desired value of 0.85, indicating no conceptual overlap. The remaining HTMT ratios (0.235-0.585) also confirm that both constructs are conceptually distinct. Generally, the model meets the conditions for discriminant validity, providing evidence of the structural integrity and separability of the constructs.

4.6. Path Coefficient

The path coefficient outcomes shown in Table 6, indicate multifaceted behavioural dynamics that go beyond the traditional Technology Acceptance Model (TAM) anticipations. The positive and statistically significant predictors are Perceived Fairness, which has strong effects on Investor Digital Literacy (β = 0.479; p < 0.001) and Robo-Advisory Adoption and Personalisation (β = 0.317; p < 0.001). This observation shows that the perception of transparency, procedural uniformity, and algorithmic impartiality are outliers in determining how investors would engage in robo-advisory systems. The environment of the retail investment in Malaysia where the elements of advisory conflict of interest and non-transparent fee frameworks still remain central issues: fairness seems to be a precondition of digital communication, rather than a secondary ethical factor.

Table 6. Path coefficients.

Note: ***Significance at 1%; **Significance at 5%; *Significance at 10%;

On the one hand, despite the classical assumptions about TAM, Perceived Usefulness does not significantly affect Investor Digital Literacy (β = 0.065; p = 0.292) and has a small impact at the 10% level of significance only (β = 0.144; p = 0.073). This paradox implies that the efficient operation or performance expectation by Malaysian investors do not necessarily transfer to behavioural acceptance. The statistically significant and positive impact of Trust in Algorithm on adoption (β = 0.197; p = 0.022). Instead of making adoption instant, increased trust can reduce the sensitivity of the investors to the authority of algorithms and the non-existence of downside risk, resulting in increased adoption of robo-advisory among investors. It is also interesting to note that trust has a positive relationship with Investor Digital Literacy (β = 0.208; p = 0.003), implying that trust inspires investors to read and investigate robo-advisory systems and determine whether to use it or not.

The mediating analysis results signify the relationship between fairness perceptions and adoption is greatly enhanced by investor Digital Literacy (β = 0.127; p = 0.015) meaning that the perceptions of fairness are converted to adoption only when investors are technologically well-endowed with the ability to adopt. Similarly, the mediating effect of Trust in Algorithm (β = 0.055; p = 0.030) indicates that trust holds mediating effects on the adoption through facilitating the capacity of investors to comprehend and consider the algorithmic advice. Conversely, the insignificant mediating effect for Perceived Usefulness validates its restricted role, strengthening that functional value alone is inadequate to determine the adoption of robo-advisory in the Malaysian context.

4.7. Model Explanatory Power

The model explanatory power is analysed through R-square and Adjusted R-square values which explains the percentage of variance in dependent that can be predicted by independent variables. The results of explanatory power are provided in Table 7.

Table 7. Explanatory power.

The results in Table 7 shows that 43.5% variation in investors digital literacy and 24% variation in rob-advisory adoption and personalisation can be predicted by trust in algorithms, perceived usefulness, and perceived fairness.

5. DISCUSSION

The H1 results leads to acceptance of the hypothesis and validates that the portfolio returns and the diversification scores significantly improved after being advised through robots. These findings are line with the Modern Portfolio Theory (MPT) that speculates that structured diversification, as opposed to single stock selection, is the cause of the best risk-return results suggested by (Ahn et al., 2020) also. The robo-advisory framework was found to adhere to MPT allocation guidelines more uniformly than the conventional static advice, which concurs with previous findings that algorithms reduce behavioural biases including home bias and familiarity bias suggested by (Boreiko & Massarotti, 2020; Khosravi, 2024) also. Nevertheless, the given enhancement can also be explained by the fact that robo-based investment suggestions make the process of making complex allocations easier and quicker than manual portfolio creation and encourage the participants to engage in the process even more, which is also suggested by (Kapur & Shrivastava, 2025) also. Also, the participants of experiment were responsive to more in the simulated environment due to the knowledge of assessment, which may have exaggerated short-term effects. In line with (Singh & Kumar, 2025), these findings are viewed as a demonstration of the potential of robo-advisory to enhance the quality of short-run allocations and diversification and not evidence of the enduring superiority of returns. To guarantee that these effects are continuing, longer-horizon and real-market applications are desirable.

The improvement in Malaysia is however more evident than other research done in technologically advanced markets where investors already have fairly diversified portfolios. By comparison, Malaysian retail investors tend to have concentrated holdings and domestic stocks or stocks familiar with the sector, which increases the effect of automatic diversification suggested by (Ajouz et al., 2025) also. Risk-aversion and low exposure to quantitative methods of portfolio management are further aspects of culture that explain why robo-advisory solutions provide better portfolio diversification score in Malaysia. Therefore, the findings substantiated the fact that algorithmic optimisation can be an effective way to improve returns and diversification, confirmed MPT, as well as emphasising significant returns to Malaysian retail investors.

The acceptance of H2 means that the level of trust in algorithms is a major factor in determining robo-advisory adoption and personalisation among the Malaysian retail investors. In line with (Ajouz et al., 2025), trust is a key antecedent of utilisation of automated financial services. Also, similar, to the consistent beneficial outcomes that were found by (Singh & Kumar, 2025) in the Indian setting, the findings in Malaysia also show a path where trust has a direct significant and positive impact on the adoption and indirectly enhances adoption through investor digital literacy. This implies that in the presence of adequate knowledge on how an algorithm works, trust will tend to increase adoption of robo-advisory, and strengthen the view of robo-advisors as an effective investment advisory tool. The investors in a risk averse culture like Malaysia are cautious about investment decisions, and financial risks, though, trust in algorithmic investment advisories with the digital literacy and expertise to assess and analyse digital advice enhance adoption and personalisation. Such findings thus, suggest that trust is also facilitated by both cognitive and digital competencies, which characterises the fact that algorithmic credibility should be augmented by investor capability and literacy also in order to achieve its practical adoption potential.

In relation to theory, trust acts as behavioural supplement to the MPT assumption on rational diversification. Whereas MPT focuses on risk-return optimisation, real investor behaviour is based on belief in the mechanism that creates portfolio recommendations. Malaysian investors will have confidence in algorithmic risk modelling and be suspicious over automated decision-making, which yield mixed outcomes. Therefore, despite the impact of trust, the impact is more significant when it is adjusted by investor literacy, that is, both cultural risk aversion and the phase of transition of the Malaysian digital finance ecosystem.

The results of H3 suggest that the perceived usefulness has a significant, though, small positive impact on the adoption and personalisation of robo-advisor in Malaysia, not as strong as those seen in other related studies. For instance, (Tahvildari, 2025) also established a further close relationship between the perceptions of usefulness and adoption of robo-advisory as the more the investors perceive that robo-advice can help increase their performance, the more they are likely to adopt it as part of an investment routine. Similarly, (Singh & Kumar, 2025) documented that the attitude and intention of the user towards the perceived usefulness have a direct impact and that it is an underlying aspect of the Technology Acceptance Model (TAM). Contrasting, the Malaysian scores indicate that usefulness has a peripheral direct impact indicating that perceived functional value alone cannot be used to ascertain behavioural uptake.

This deviation may be due to cultural and behavioural factors. For instance, in Malaysian retail investor tends to use a human financial adviser or peer network, and inter-personal validation is more important than functional benefits of the system (Chen et al., 2025). As a result, usefulness cannot be understood as a driver as it would only be meaningful when combined with trust and literacy. Relative to MPT, usefulness is consistent with the rational optimisation of risk-reward portfolio by investors; nonetheless, unless the benefits of algorithmic optimisation are made clear to them, Malaysian investors might not consider algorithmic optimisation useful. Consequently, usefulness is important, but its effects are conditional, and not so decisive as trust and fairness.

The findings of H4 are also accepted and support the claim that the perceived fairness is a significant predictor of robo-advisory adoption and personalisation, which is in line with (Kapur & Shrivastava, 2025) who highlighted that the perceived transparency of algorithms and unbiased advice lie at the centre of investor-related confidence towards conversational AI and robo-advisors. Fairness seems to be even more of a determinant than usefulness in Malaysia implying that ethical and distributional issues play a significant role in decision-making in investments. It is in contrast with the Chinese setting in (Chen et al., 2025) where the driving forces of adoption and personalisation and ESG characteristics prevailed, whereas the issues of fairness were secondary. The more sensitive nature of Malaysians to fairness can be explained by the cultural focus on procedural justice, the increased sensitivity to financial misconduct scandals, and the worries about the lack of transparency in algorithms.

In the MPT perspective, fairness does not affect directly on the optimisation of risk-return, it determines the willingness to trust such risk-return optimisation systems. Malaysian investors appear to be unwilling to pass the control of the portfolio to algorithms unless they believe that the recommendations provided are not biased, conflict-free, and do not have any hidden commercial objectives revealed by (Muganda & Kasamani, 2023; Rossi & Utkus, 2020) also. This implies that there is mediation of rational adoption by MPT in terms of behavioural norms. Thus, the perceived fairness serves as both a psychological protective mechanism and a behaviour change agent, which confirms that transparent algorithms are significant to enhancing adoption and personalisation for diversification of portfolios within the new robo-advisory environment in Malaysia.

The validation of H5 evidently signifies investor digital literacy as a significant mediating factor to transform behavioural insights such as trust, perceived usefulness, and fairness into robo-advisory adoption and personalisation. (Chen et al., 2025) also defines digital literacy as a flexible phenomenon, though the concept is typically viewed as an exogenous capacity that determines the use of technology, whereas the study describes it as such that can transform during the interaction with algorithmic systems suggested in the findings of (Ajouz et al., 2025) as well. In the experimental setting, exposure to personalised robo-advisory recommendations has the potential to promote the knowledge of system logic in participants and thus, perceived digital competence. This rationale corresponds to the literacy as a post-exposure cognitive adaptation and not a trait. Instead of being a direct performance driver, literacy mediates the process of cognition of trust, perceived usefulness, and perceived fairness in processing and acting, which is consistent with (Ajouz et al., 2025), who opines that algorithmic awareness can only reduce uncertainty when investors can meaningfully interpret system outputs. The same is reflected by (Singh & Kumar, 2025) who demonstrate that the attitudinal effects are enhanced by digital familiarity, but the Malaysian data indicates a steeper learning curve whereby exposure to AI finance is limited at the outset which limits the uptake of behaviour. The difference with (Chen et al., 2025), who use mediation, could also be due to the differences in the context, such as, in Malaysia, the digital competencies are heterogeneous, so trust or fairness signals cannot be used on their own without the competencies to assess them.

Another specification could position digital literacy as an antecedent instead of an outcome of personalised exposure to robo-advisory, which raises the possibility of endogeneity concerns in a single-wave structural equation model. It is probable that more digitally literate investors in the past are more inclined to gain, accept or even make sense of robo-advisory advice which was unveiled in the discussion of (Singh & Kumar, 2025) as well. In this study, though, digital literacy was assessed as perceived competence after experimental exposure to personalised robo-advisory situations. The directional specification was temporally plausible since the quasi-experimental design guaranteed that the participants were exposed to the robo-advisory simulation first and then surveyed using the survey instrument. Literacy is, therefore, a post exposure evaluative competence and not pre-existing ability which is indicated by (Kapur & Shrivastava, 2025) also. However, the prospective studies are suggested to investigate bidirectional or longitudinal models because the prospective study could test the potential of bidirectional relationships that might exist between digital literacy, adoption behaviour, and portfolio performance in the long-term.

The effects of learning-curve can be also exacerbated by short-term experimental exposure, in which case digitally skilled individuals are advantaged by the learning-curve, and less literate investors are either wary or do as expect. This means that the results cannot be viewed as evidence of the improved robo-advisory results. In terms of MPT, literacy increases the capacity of an investor to be aware of diversification algorithms and risk-optimised allocations in the gap between theoretical optimisation and behavioural acceptance. Therefore, digital literacy is the means of converting abstract trust and subjective beliefs of fairness into the adoption intentions. The findings highlight the importance of improving digital capacities of Malaysian investors to enable maximum attainment of the benefits of robo-advisory.

CONCLUSION

The study demonstrates that the adoption of robo-advisory in Malaysia is not influenced by the technology but a multifaceted interaction of behavioural, ethical, and capability-based characteristics. Although algorithmic models have been promised to ensure optimisation in line with MPT, the implied tools are only used by Malaysian investors when trust, fairness, and literacy factors are satisfied- a continued disconnecting the gap between the theoretical and actual efficiency. The experiment also demonstrates that the returns and diversification are improved using algorithmic advice, but the behavioural constraints restrict its implementation on a broad basis. This shows that an increase in technological development does not necessarily lead to empowerment of investors’ return and investment decisions, unless they rely equally on literacy and governance. Generally, the study states that the potential of robo-advisory in Malaysia is depended on algorithmic transparency, trust, investor digital literacy, and culture-sensitive trust-building processes will follow the technology.

LIMITATIONS & FUTURE DIRECTION

The study while holds valuable empirical and theoretical contributions, it also holds certain limitations as well. For instance, at first, the study relied on 30-days simulated investment period which possible does not capture market volatility over the long-run period or behavioural adaptation of investors towards robo-advisory tools. Furthermore, the sample is sufficient but there are geographical and demographical limitation and this limits generalisability of the sample to the wider investor population in Malaysia. Bias also be brought about by self-reported survey responses. Future studies need to include longitudinal real-market data, increase diversity of participants and also incorporate aspects of qualitative research to explore the nuances of behaviour in a deeper way. The comparison of studies between ASEAN markets and the experiment with various robo-advisory services can be helpful to better figure out the contextual factors and enhance the external validity of investment advice provided by algorithms. Lastly, the major limitation of the research is that it utilises short-term, simulated portfolio returns, thus limiting the ability to make inferences on the long- term performance, risk behaviour, and market-cycle effects. The short run returns can be due to short-term adjustments or learning as opposed to value creation. To determine the persistence, drawdowns and behavioural adjustment to robo-advisory systems over time, future studies need to utilise longitudinal designs with real accounts of investments or protracted simulators across different market conditions.

IMPLICATIONS

The results provide tentative policy suggestions to the changing digital investment environment in Malaysia as opposed to prescriptive ones. Investigators like the Securities Commission Malaysia can view these findings as the initial evidence pointing to the significance of usefulness, algorithmic trust, and communication and transparency for retail investors to guarantee enhanced robo investment advisory services and its adoption. The insights can be used by financial institutions and fintech providers to enhance pilot-level adoption of robo-advisory services especially among digitally literate retail investors. Moreover, the investor education agencies can consider that digital literacy as a facilitating provision that contributes to informed usage of automated advisory tools. In general, the research feeds the current policy discussion of responsible fintech experimentation, though it does not suggest immediate massive regulation and market reorganization. The implications of the present study also comprise theoretical and practical implications. Hypothetically, the results broaden the perspectives of technology acceptance and behavioural finance by underscoring the ascendancy of perceived fairness as well as the situational role of digital literacy in the algorithm-driven investment contexts. In practise, the findings guide fintech developers and financial institutions to increase the level of transparency, credibility, and interaction with users in robo-advisory systems. Policymakers are also suggested to look to encourage algorithmic disclosure and investor education as a way of adopting them responsibly. All these implications collectively show that personalised robo-advisory systems are driving investment behaviour, portfolio optimisation, and participation of investors in emerging digital finance ecosystems.

LIST OF ABBREVIATIONS

AUTHOR’S CONTRIBUTION

I.A.B. has contributed to conceptualization, idea generation, problem statement, methodology, results analysis, results interpretation.

ETHICAL STATEMENT & INFORMED CONSENT

All procedures were conducted in compliance with the guidelines of the institutional research ethics committee and adhered to the principles outlined in the Declaration of Helsinki. Informed consent was obtained from all participants prior to their inclusion in the study. To protect participant confidentiality, all data were anonymized at the time of collection, and no personally identifiable information was recorded.

AVAILABILITY OF DATA AND MATERIALS

The data will be made available on reasonable request by contacting the corresponding author [I.A.B.].

FUNDING

None.

CONFLICT OF INTEREST

The author declares no conflicts of interest, financial or otherwise.

ACKNOWLEDGEMENTS

Declared none.

DECLARATION OF AI

During the preparation of this manuscript, the author used ChatGPT for language polishing. After utilizing this tool, the author carefully reviewed and refined the content as necessary and accept full responsibility for the accuracy and integrity of the published work.

APPENDIX A

Demographic Factors

Survey Questionnaire

REFERENCES

Abbas, S. K. (2024). AI meets finance: the rise of AI-powered Robo-advisors. Journal of Electrical Systems, 20(11), 1011-1016. https://doi.org/10.52783/jes.7359

Adegbenro, A., Madueke, A., Ojo, A., & Alabi, C. (2022). AI-Driven Wealth Advisory: Machine Learning Models for Personalized Investment Portfolios and Risk Optimization. Communication In Physical Sciences, 8(4), 779-807. Available from: https://journalcps.com/index.php/volumes/article/view/739

Adewale, T. (2025). The Rise of Algorithmic Financial Advice: Robo-Advisors, Fintech Synergies, and Investment Theory in Practice. International Journal of Research Publication and Reviews, 6(1), 4832-4848. https://doi.org/10.55248/gengpi.6.0125.0640

Afifah, S., Mudzakir, A., & Nandiyanto, A. B. D. (2022). How to calculate paired sample t-test using SPSS software: From step-by-step processing for users to the practical examples in the analysis of the effect of application anti-fire bamboo teaching materials on student learning outcomes. Indonesian Journal of Teaching in Science, 2(1), 81-92. https://doi.org/10.17509/ijotis.v2i1.45895

Ajouz, M., Abu-AlSondos, I. A., Yaseen, S. G., Alkhwaldi, A. F., Al Moghrabi, A., & Al-Qaisi, N. (2025). Portfolio Management Through Financial Robo-advisors: Key Drivers of Potential Investor Adoption. In Applied Artificial Intelligence in Business: Systems, Tools and Techniques (pp. 283-294). Cham: Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-90271-0_21

Akhtar, F., Akhtar, S., & Laeeq, M. (2025). Evolution of Robo‐Advisors: A Literature Review and Future Research Agenda. International Journal of Consumer Studies, 49(6), e70131. https://doi.org/10.1111/ijcs.70131

Ali, A., Salleh, M. A. M., & Mustaffa, N. (2022). Digital inheritance exploration (DIE) through digital inheritance model (DIM)—A guideline for digital assets planning after death. APS proceedings. Academica Press Solution. Available from: https://www.researchgate.net/publication/365315619 (Accessed on: May 12, 2025).

Aw, E. C. X., Leong, L. Y., Hew, J. J., Rana, N. P., Tan, T. M., & Jee, T. W. (2024). Counteracting dark sides of robo-advisors: justice, privacy and intrusion considerations. International Journal of Bank Marketing, 42(1), 133-151. https://doi.org/10.1108/IJBM-10-2022-0439

Ballance, O. J. (2024). Sampling and randomisation in experimental and quasi-experimental CALL studies: Issues and recommendations for design, reporting, review, and interpretation. ReCALL, 36(1), 58-71. https://doi.org/10.1017/S0958344023000162

Bhattacharjee, J., Singh, R., Pandey, S., & Sharma, R. (2025). Factors impacting stress in financial investment due to the use of artificial intelligence: a social networking analysis approach. AI & SOCIETY, 1-14. https://doi.org/10.1007/s00146-025-02754-4

Boreiko, D., & Massarotti, F. (2020). How risk profiles of investors affect robo-advised portfolios. Frontiers in Artificial Intelligence, 3, 60. https://doi.org/10.3389/frai.2020.00060

Capponi, A., Olafsson, S., & Zariphopoulou, T. (2022). Personalized robo-advising: Enhancing investment through client interaction. Management Science, 68(4), 2485-2512. https://doi.org/10.1287/mnsc.2021.4014

Chan, S. C., Liew, X. W., Lim, J. Y., & Man, S. K. (2023). Determinants of the intention to use financial robo-advisory in Malaysia. UTAR Institutional Repository. http://eprints.utar.edu.my/6059/1/fyp_FN_2023_CSC.pdf

Chen, A., Wang, S., Mehta, A. M., Asif, M., Xu, S., & Shahzad, M. F. (2025). FinTech adoption for ESG integration through robo advisors, personalization, and perceived trust. Scientific Reports, 15(1), 31125. https://doi.org/10.1038/s41598-025-17046-6

Derbali, A., & Lamouchi, A. (2020). Global financial crisis, foreign portfolio investment and volatility: Impact analysis on select Southeast Asian markets. Pacific Accounting Review, 32(2), 177-195. https://doi.org/10.1108/PAR-07-2019-0090

Dias, F. S. (2025). An Introduction to Robo-Advisors, Notable Fintech Implementations, and Underlying Theory. In Fintech and the Emerging Ecosystems: Exploring Centralised and Decentralised Financial Technologies (pp. 285-300). Cham: Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-83402-8_11

Fantazy, K., & Tipu, S. A. A. (2024). Linking big data analytics capability and sustainable supply chain performance: mediating role of knowledge development. Management Research Review, 47(4), 512-536. https://doi.org/10.1108/MRR-01-2023-0018

George, A. S. (2024). Robo-revolution: Exploring the rise of automated financial advising systems and their impacts on management practices. Partners Universal Multidisciplinary Research Journal, 1(4), 1-6.

Haji-Othman, Y., & Yusuff, M. S. S. (2022). Assessing reliability and validity of attitude construct using partial least squares structural equation modeling. Int J Acad Res Bus Soc Sci, 12(5), 378-385. https://doi.org/10.6007/IJARBSS/v12-i5/13289

Jailani, M. N. A., & Adenan, F. (2023). Shariah Risk Management In Islamic Digital Banking In Malaysia: A Study At KAF Investment Bank: Pengurusan Risiko Syariah Dalam Perbankan Digital Islam Di Malaysia: Kajian Di KAF Investment Bank. al-Qanatir: International Journal of Islamic Studies, 30(2), 188-197. Available from: https://al-qanatir.com/aq/article/view/678

Kapur, G., & Shrivastava, S. (2025). Evaluating the Role of Conversational AI in Financial Investment Decision Making in NCR. International Journal of Economic Practices and Theories, 247-262. https://doi.org/10.52783/ijept.35

Khosravi, F. (2024). Transforming Investment Advisory Services Through Artificial Intelligence: A Study on Robo-Advisors and Algorithmic Portfolio Management. Nuvern Applied Science Reviews, 8(9), 1-8. https://nuvern.com/index.php/nasr/article/view/2024-09-04

Koneti, S. B. (2025). Artificial intelligence Applications in Retail and Investment Banking: Personalization, Robo-Advisory and Behavioral Analytics. Artificial Intelligence-Powered Finance: Algorithms, Analytics, and Automation for the Next Financial Revolution, 4, 72. https://doi.org/10.70593/978-93-7185-613-3

Kulkarni, M. S., Patil, K. P., & Pramod, D. (2025). The role of robo-advisors in behavioural finance, shaping investment decisions. Cogent Economics & Finance, 13(1), 2571403. https://doi.org/10.1080/23322039.2025.2571403

Kumar, K. R., Balaji, C., Suresh, M., Nagaraju, E., Natarajan, S., & Kumar, R. (2024). Study The Impact Of Ai Driven Rob Advisors on Wealth Management Services. Evaluate Their Effectiveness in Portfolio Management, Risk Assessment, And Personalized Investment Strategies, 30(5), 5819. https://doi.org/10.53555/kuey.v30i5.5819

Mahdzan, N. S., Zainudin, R., & Yoong, S. C. (2020). Investment literacy, risk tolerance and mutual fund investments: An exploratory study of working adults in Kuala Lumpur. International journal of Business and Society, 21(1), 111-133. https://doi.org/10.33736/ijbs.3230.2020

Mathur, P., & Sharma, A. M. (2024). Robo-advisors in investment management. In Risks and challenges of AI-driven finance: Bias, ethics, and security (pp. 121-145). IGI Global Scientific Publishing. https://doi.org/10.4018/979-8-3693-2185-0.ch006

MDEC. (2025). Available from: https://mdec.my/media-release/news-press-release/375/malaysia%E2%80%99s-digital-investments-hit-record–rm163.6-billion-in-2024 (Accessed on: November 11, 2025).

Memon, M. A., Ramayah, T., Cheah, J. H., Ting, H., Chuah, F., & Cham, T. H. (2021). PLS-SEM statistical programs: a review. Journal of Applied Structural Equation Modeling, 5(1), 1-14. https://doi.org/10.47263/JASEM.5(1)06

Mohd Dzin, N. H., & Lay, Y. F. (2021). Validity and reliability of adapted self-efficacy scales in Malaysian context using PLS-SEM approach. Education Sciences, 11(11), 676. https://doi.org/10.3390/educsci11110676

Puspaningtyas, M. (2022). The Role of Technology and Investment in Digitalization and The Economy in Malaysia. Tamansiswa Management Journal International, 4(1), 35-35. https://doi.org/10.54204/TMJI/Vol412022006

Qing, S., & Valliant, R. (2025). Extending Cochran’s sample size rule to stratified simple random sampling with applications to audit sampling. Journal of Official Statistics, 41(1), 309-328. https://doi.org/10.1177/0282423X241277054

Roh, T., Park, B. I., & Xiao, S. S. (2023). Adoption of AI-enabled Robo-advisors in Fintech: Simultaneous Employment of UTAUT and the Theory of Reasoned Action. Journal of Electronic Commerce Research, 24(1), 29-47.

Russo, G. M., Tomei, P. A., Serra, B., & Mello, S. (2021). Differences in the use of 5-or 7-point likert scale: an application in food safety culture. Organizational Cultures, 21(2), 1. https://doi.org/10.18848/2327-8013/CGP/v21i02/1-17

Securities Commission Malaysia. (2025). Available from: https://www.sc.com.my/resources/media/media-release/malaysian-capital-market-hits-record-rm42-trillion-in-2024-stays-resilient-amid-earnings-growth (Accessed on: November 11, 2025).

Singh, S., & Kumar, A. (2025). Investing in the future: an integrated model for analysing user attitudes towards Robo-advisory services with AI integration. Vilakshan-XIMB Journal of Management, 22(1), 158-175. https://doi.org/10.1108/XJM-03-2024-0046

Syed, W. K., & Janamolla, K. R. (2024). How AI-driven robo-advisors impact investment decision-making and portfolio performance in the financial sector: A comprehensive analysis. Int. Res. J. Eng. Technol, 11, 138-145.

Tahvildari, M. (2025). Analysis of investor profiling and portfolio recommendations by Robo-Advisors in Germany. Global Business & Finance Review, 30(8), 19. https://doi.org/10.17549/gbfr.2025.30.8.19

Yang, Q., & Lee, Y. C. (2024). Enhancing financial advisory services with GenAI: Consumer perceptions and attitudes through service-dominant logic and artificial intelligence device use acceptance perspectives. Journal of Risk and Financial Management, 17(10), 470. https://doi.org/10.3390/jrfm17100470

Yi, T. Z., Rom, N. A. M., Hassan, N. M., Samsurijan, M. S., & Ebekozien, A. (2023). The adoption of robo-advisory among millennials in the 21st century: trust, usability and knowledge perception. Sustainability, 15(7), 6016. https://doi.org/10.3390/su15076016