Abstract
Artificial intelligence-enabled financial technology is reshaping the ways in which small and medium-sized enterprises access, use, and evaluate formal financial services. The theoretical problem addressed in this study is that firms with comparable levels of AI-enabled FinTech adoption may achieve substantially different levels of business financial inclusion. This inconsistency suggests that adoption creates a technological opportunity structure, but does not by itself ensure effective access, use, quality, or value creation. Digital financial capability is therefore positioned as the capability-conversion mechanism through which SMEs transform AI-enabled financial tools into meaningful financial participation. This study identifies the configurational pathways through which AI-enabled FinTech adoption and digital financial capability are associated with high business financial inclusion among small and medium-sized enterprises in northern Peru. A quantitative, non-experimental, cross-sectional design was applied to 120 complete responses from owners, managers, and administrators with direct knowledge of their firms’ financial and digital operations. The analysis used fuzzy-set qualitative comparative analysis, incorporating six causal conditions: AI-enabled FinTech adoption, digital financial capability, perceived usefulness, AI familiarity and integration, digital financial skills, and digital financial self-efficacy. The findings show that no single condition is strictly necessary for high business financial inclusion. Nevertheless, digital financial capability emerged as the condition closest to necessity. The intermediate solution identified two equifinal configurations leading to high inclusion, both combining AI-enabled FinTech adoption, digital financial capability, perceived usefulness, and digital financial skills, while AI familiarity and digital financial self-efficacy operated as alternative reinforcing mechanisms. The results also confirmed causal asymmetry, as low inclusion was explained by configurations centred on deficits in capability and AI-FinTech adoption rather than by the simple inverse of high-inclusion pathways. The study extends FinTech and financial inclusion theory by demonstrating that business financial inclusion is not a direct result of AI-enabled FinTech adoption, but an outcome of capability conversion operating through multiple configurational pathways.
1 Introduction
The convergence between financial technology services and artificial intelligence has reconfigured the architectures of intermediation, credit assessment, and digital payments, establishing a paradigm in which machine-learning algorithms and conversational assistants operate as a core layer of the financial offering (; ). Against this backdrop, the Global Findex reports that 79% of adults worldwide have a financial or mobile money account, compared with 51% in 2011 (World Bank, 2025); Latin America and the Caribbean reached 70% in 2024, with mobile money increasing from 22% to 37% between 2021 and 2024 (; World Bank, 2025). Such expansion, however, does not automatically translate into business inclusion, because the effective use of AI-based FinTech services depends on the combination of digital financial knowledge, operational skills, and self-efficacy (; ; ).
In the Peruvian context, these tensions become particularly clear. Although sustained progress in physical and digital coverage was recorded between 2019 and 2024 (Superintendencia de Banca, Seguros y AFP, 2025), 33% of the adult population remains outside the formal financial system (). At the same time, the Peruvian FinTech ecosystem reached 193 local ventures and 153 foreign ventures by the end of 2024, with the exit rate decreasing from 17% to 12% in the most recent cycle (), while the business sector closed 2024 with 2,346,592 formal firms, of which 99.1% were micro and small enterprises that contributed 20.2% of GDP and accounted for 86.5% of private employment (). The northern region of the country, comprising La Libertad, Lambayeque, and Piura, concentrates a strategic SME density, since La Libertad and Piura alone account for 9.4% of national formal MSEs and combine high mobile-banking penetration with low business digital literacy (). This makes the region an optimal empirical setting for examining the pathway from AI-enabled FinTech adoption to business financial inclusion.
Recent international evidence has advanced the understanding of the individual antecedents of FinTech adoption. demonstrated, through structural equation modelling, that digital financial literacy operates as both predictor and mediator, whereas financial capability and autonomy emerge as dominant mediators; confirmed the mediating role of digital financial literacy under the moderation of perceived trust; and extended the technology acceptance model by incorporating familiarity with artificial intelligence as a direct antecedent, evidence later replicated among female users by . At the SME level, Setiawan et al. (2025, 2026) identified urban-rural differentiation in the mediation of digital financial literacy; Tandilino et al. (2025) reported that digital financial inclusion mediates the relationship between digital financial literacy and business performance; confirmed the mediation of digital payment adoption; and and documented differentiated effects by gender and territory. These findings are consistent with Zhang and Fan (2024), Zaimovic et al. (2025), and , who showed that financial knowledge translates into effective use only when accompanied by operational skills and digital self-efficacy.
Additionally, , , and Tetteh and Kwateng (2026) consolidated the behavioural perspective by reframing digital capability and self-efficacy as primary drivers of well-being and sustained engagement with mobile financial services, while , , Singh et al. (2025), and provided complementary evidence on the socioeconomic effects of inclusion and on capability expansion as a pathway to inclusion. Although most approaches use structural equation modelling under linear assumptions, an emerging stream has incorporated fuzzy-set qualitative comparative analysis as a complementary strategy; in this direction, , , and demonstrated the relevance of identifying sufficient configurations and causal asymmetries in the FinTech domain, although such applications have concentrated on Asian and sub-Saharan contexts, without documented replication in Latin American SMEs.
Based on the evidence reviewed, the central gap addressed by this study is not merely the scarcity of Peruvian evidence or the limited use of fsQCA in Latin American FinTech research. The more fundamental theoretical problem is the persistent inconsistency between technological adoption and inclusion outcomes: SMEs may report similar levels of AI-enabled FinTech adoption, yet differ substantially in their access to formal financial products, their regular use of digital financial services, and the perceived quality of those services. This inconsistency challenges linear adoption-based explanations and suggests that AI-enabled FinTech adoption should be understood as an opportunity-generating condition rather than as a sufficient cause of financial inclusion. From this perspective, digital financial capability becomes the central conversion mechanism that determines whether firms can transform algorithmic credit assessment, predictive analytics, automated recommendations, and AI-assisted financial interfaces into effective financial participation and business value. The contextual relevance of northern Peru and the methodological relevance of configurational analysis are therefore treated as boundary conditions that allow this theoretical problem to be examined in an underexplored Andean SME setting, rather than as claims that fsQCA, the constructs, or the relationships are entirely new. Evidence from rural settings likewise indicates that FinTech adoption can strengthen financial inclusion when contextual and capability constraints are addressed ().
From this framing, the present study poses the following general research question: how does AI-enabled FinTech adoption articulate with business financial inclusion among small and medium-sized enterprises in northern Peru, considering the central role of digital financial capability in the sufficient configurations for high inclusion? Three specific questions derive from this general problem. First, the study examines in which configurations sufficient for high business financial inclusion AI-enabled FinTech adoption appears together with digital financial capability. Second, it explores in which configurations sufficient for high business financial inclusion digital financial capability appears. Third, it asks whether digital financial capability operates as a central condition that articulates the presence of AI-enabled FinTech adoption in the configurations sufficient for high business financial inclusion.
Accordingly, the general objective is to identify the sufficient configurations of AI-enabled FinTech adoption and digital financial capability associated with business financial inclusion among small and medium-sized enterprises in northern Peru, examining the central role of digital financial capability within those configurations. Three specific objectives structure the analytical pathway: the first is to examine the configurational role of AI-enabled FinTech adoption and of its dimensions of perceived usefulness and familiarity with artificial intelligence as conditions present in the configurations sufficient for high business financial inclusion; the second is to analyse the configurational role of digital financial capability and of its dimensions of digital financial skills and digital financial self-efficacy as conditions present in the configurations sufficient for high business financial inclusion; and the third is to determine whether digital financial capability operates as a central condition that articulates the presence of AI-enabled FinTech adoption in the configurations sufficient for high inclusion, as well as to contrast causal asymmetry between the configurations associated with high and low inclusion.
Consistent with the configurational approach, four hypotheses are formulated in terms of the joint presence of conditions, centrality, and causal asymmetry (; ; ; ). The first hypothesis states that AI-enabled FinTech adoption and its dimensions of perceived usefulness and familiarity with artificial intelligence appear as conditions present in the configurations sufficient for high business financial inclusion among SMEs in northern Peru. The second hypothesis posits that digital financial capability and its dimensions of digital financial skills and digital financial self-efficacy also appear as conditions present in such sufficient configurations. The third hypothesis states that digital financial capability operates as a central condition, concurrently with AI-enabled FinTech adoption, in the configurations sufficient for high business financial inclusion. Finally, the fourth hypothesis proposes that the configurational explanation of low business financial inclusion does not constitute the inverted symmetrical image of the explanation of high inclusion, thereby evidencing causal asymmetry in the sense proposed by .
The study offers theoretical, methodological, empirical, contextual, and practical contributions. Theoretically, it extends technology acceptance, capability, and financial inclusion perspectives by proposing a capability-conversion explanation: AI-enabled FinTech adoption creates digital financial opportunities, digital financial capability converts those opportunities into firm-level action, and business financial inclusion emerges as the outcome of that conversion. This formulation refines prior adoption-centred explanations by showing that the relevant theoretical issue is not only whether firms accept AI-enabled FinTech, but whether they possess the cognitive, operational, and self-efficacy resources required to convert adoption into inclusion. Methodologically, the study does not present fsQCA as a new method; rather, it applies a configurational logic to a phenomenon that is theoretically characterized by equifinality, causal complexity, and asymmetry. Empirically and contextually, it provides evidence from Peruvian Andean SMEs, an underexplored Latin American setting with high SME density and uneven digital financial literacy. Practically, it provides inputs for regulators, business associations, and FinTech providers by identifying which combinations of adoption, usefulness, skills, capability, AI familiarity, and self-efficacy should be strengthened to move firms from technological access to effective financial inclusion. The study is also aligned with the financial inclusion and innovation objectives of the 2030 Agenda for Sustainable Development (United Nations, 2015).
2 Theoretical framework
AI-enabled FinTech adoption is conceived as a behavioural and organizational process through which economic agents incorporate and use financial services whose operational core is enhanced by artificial intelligence. Unlike traditional digital financial services, AI-enabled FinTech incorporates algorithmic decision making, automated credit assessment, predictive analytics, intelligent recommendation systems, anomaly and fraud detection, and AI-assisted financial decision interfaces. Related research has examined AI-enabled voice assistants, customer-facing AI financial services, explainable credit assessment, and central bank digital currency adoption as additional expressions of AI-enabled financial services (; ; ; ). These features alter the nature of adoption because firms are not only using a digital channel; they are interacting with systems that classify risk, anticipate needs, personalize offers, automate advice, and influence financial decisions. Consequently, the theoretical contribution of the AI component lies in explaining how algorithmic functionalities expand the opportunity space available to SMEs while also increasing the need for interpretive and operational capabilities.
Building on TAM and UTAUT (; Venkatesh et al., 2003, 2012), perceived usefulness, ease of use, facilitating conditions, and continued-use intention explain why firms become willing to adopt AI-enabled FinTech. Recent evidence also supports these mechanisms in AI familiarity, contactless-payment continuance, trust-mediated robo-advisor adoption, and intention to recommend AI in financial services (; ; ; ; ). However, these theories do not fully explain why similar levels of adoption may produce different inclusion outcomes. To address this limitation, this study integrates technology acceptance theory with the capability approach and financial inclusion theory. In the integrated framework, AI-enabled FinTech adoption represents the creation of technological opportunities; digital financial capability represents the conversion process that enables firms to understand, evaluate, operate, and appropriate those opportunities; and business financial inclusion represents the realized outcome in terms of access, use, quality, and satisfaction with formal financial services.
Digital financial capability is therefore defined not as an additional antecedent placed beside technology adoption, but as the core explanatory mechanism that links adoption to inclusion. Following the capability approach (Sen, 1999), inclusion depends on whether firms can convert available resources into valued functionings. In AI-enabled financial environments, this conversion requires knowledge of digital products and algorithmic risks, operational skills to execute and monitor transactions, financial behaviours that translate information into disciplined decisions, and self-efficacy to evaluate recommendations, correct errors, and use services autonomously. Thus, a firm may adopt AI-enabled FinTech and still fail to achieve meaningful financial inclusion if it lacks the capability to convert technological access into effective financial participation and value creation. Related evidence links financial advice seeking, distrust of financial institutions, financial confidence, household financial capability, and financial anxiety to financial behavior and well-being (; ; Sajid et al., 2024; Sun et al., 2022; Xiao and Meng, 2024).
From the articulation of the foregoing frameworks, four operational dimensions of digital financial capability are derived. The first, digital financial knowledge, comprises the understanding of digital financial concepts, products, and processes, as well as types of services, contractual terms, interest rates, fees, and algorithmic risks; documented, in a study of 6,200 young adults, that financial knowledge operates as a significant antecedent of well-being mediated by financial capability. The second dimension, digital financial skills, encompasses the operational abilities required to use platforms, navigate applications, execute transactions, and manage authentication; confirmed that these skills are robust predictors of financial inclusion among Indonesian millennials. The third dimension, digital financial behaviour, reflects the frequency and consistency of healthy practices carried out in digital environments, such as planning, recording transactions, and monitoring expenses; showed, in a sample of Turkish mutual fund investors, that behaviour links literacy with risk tolerance. The fourth dimension, digital financial self-efficacy, refers to conception of the belief in one's own ability to perform behaviours; consolidated, through a bibliometric analysis of 1,247 documents, that self-efficacy is a critical determinant of financial behaviour in AI-assisted environments.
Business financial inclusion is conceptualized as the realized firm-level outcome of this conversion process. It involves effective access to, regular use of, and perceived quality of formal financial services, such as savings, credit, payments, insurance, and investment, under conditions of cost, convenience, transparency, and protection appropriate to operational needs. Within the integrated framework, access alone is insufficient: inclusion becomes meaningful only when AI-enabled services are used recurrently, understood by the firm, trusted sufficiently for business decisions, and evaluated as useful for financial management. This view connects financial inclusion theory with capability theory by treating inclusion as an achieved functioning rather than as a direct consequence of digital availability.
Despite documented progress in the conceptualization of the three constructs, the relationship between AI-enabled FinTech adoption, digital financial capability, and business financial inclusion is theoretically configurational. The phenomenon involves equifinality because different combinations of AI familiarity and self-efficacy may support similar inclusion outcomes; causal complexity because adoption, usefulness, skills, and capability operate jointly rather than independently; and asymmetry because the absence of inclusion is not expected to be the simple inverse of inclusion. Accordingly, fsQCA is not used as a methodological preference, but as an analytical strategy aligned with the theoretical structure of the phenomenon. Linear approaches are useful for estimating average net effects, but they are less suitable for identifying sufficient recipes through which SMEs convert AI-enabled technological opportunities into financial inclusion. The phenomenon involves equifinality, causal complexity, and asymmetry (; ). Comparable configurational applications have also examined cryptocurrency adoption and gamified mobile-wallet use (; Yang et al., 2023).
3 Materials and methods
This study adopted a quantitative, non-experimental, cross-sectional, and explanatory-correlational design. The target population consisted of small, and medium-sized enterprises operating in northern Peru, specifically in La Libertad, Lambayeque, and Piura, and the empirical unit of analysis was the firm represented by one owner, manager, or administrator with direct knowledge of its financial and digital operations. The population was defined in substantive terms as formally or operationally active SMEs that use, evaluate, or could reasonably adopt digital financial services for payments, collections, credit, savings, or financial management. Official statistics indicate that Peru closed 2024 with 2,346,592 formal firms, of which 2,326,126 were formal micro and small enterprises, representing 99.1% of the business structure (). However, no complete, updated, and publicly accessible firm-level sampling frame with contact data for all eligible SMEs in the three study regions was available to the researchers. For this reason, the study used a non-probability purposive sampling strategy, complemented by convenience access through business and professional networks. Participants were eligible when they were owners, managers, or administrators of SMEs located in the three selected regions, had direct knowledge of the firm's financial and digital operations, provided voluntary participation, and completed the questionnaire. Responses were excluded when the respondent was outside the target regions, did not occupy a role connected with financial or digital decision making, submitted duplicate information, or left incomplete questionnaire blocks. Because recruitment did not generate a closed denominator of all eligible firms invited to participate, a conventional response rate cannot be calculated. Consequently, the final sample of 120 complete responses should be interpreted as an exploratory and analytically suitable sample for configurational modelling, not as a statistically representative sample of all SMEs in northern Peru.
Information was collected through a structured questionnaire using a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree. The instrument was designed to measure three core theoretical constructs: AI-enabled FinTech adoption, digital financial capability, and business financial inclusion. AI-enabled FinTech adoption was measured through 18 items grouped into four dimensions: perceived usefulness of AI-FinTech services, perceived ease of use, familiarity and integration with artificial intelligence, and continued-use intention and behaviour. Perceived usefulness captured the extent to which firms considered AI-FinTech services useful for improving financial performance, transactional efficiency, decision quality, time management, and predictive financial information. Perceived ease of use assessed interface clarity, learning effort, availability of support, and the intuitive use of automated functionalities. Familiarity and integration with artificial intelligence assessed the recognition of AI-based functions, understanding of algorithmic recommendations, use of chatbots, trust in automated credit assessment, and perceived value of anomaly or fraud detection systems. Continued-use intention and behaviour measured prospective intention, frequency of use, recommendation to other SMEs, and willingness to explore new AI-based functionalities.
Digital financial capability was assessed through 16 items distributed across four dimensions: digital financial knowledge, digital financial skills, digital financial behaviour, and digital financial self-efficacy. Digital financial knowledge referred to the understanding of FinTech products, contractual terms, fees, algorithmic risks, and consumer rights in digital financial environments. Digital financial skills captured the operational abilities associated with platform navigation, digital transactions, authentication, and basic troubleshooting. Digital financial behaviour assessed digital planning, systematic recording of financial transactions, reconciliation practices, and the use of analytical tools for monitoring business expenses. Finally, digital financial self-efficacy measured confidence, autonomy, critical evaluation of algorithmic recommendations, the ability to correct transaction errors, and independent management of FinTech services.
Business financial inclusion was operationalized through 13 items organized into three dimensions: access to digital financial services, use of digital financial services, and quality and satisfaction. The access dimension included the availability of bank and digital accounts, digital credit products, electronic payment platforms, as well as the adequacy of access costs and requirements. The use dimension measured the regularity and diversity of digital financial practices, including electronic payments, digital collections, bank transfers, digital financing, and digital savings or investment platforms. The quality and satisfaction dimension captured the perceived fit with business needs, transparency of information, protection mechanisms, and overall satisfaction with digital financial services.
Given that all variables were collected from a single respondent per firm through a self-report questionnaire, common method variance was explicitly considered. Procedural remedies included voluntary participation, anonymity of responses, neutral wording of items, and separation of the questionnaire into construct blocks to reduce evaluation apprehension and automatic consistency in responses. As a diagnostic remedy, an unrotated single-factor test was conducted on the available composite indicators. Because no theoretically unrelated marker variable was included in the instrument, a marker-variable test could not be applied; this limitation is acknowledged and future studies should incorporate marker variables, multi-respondent designs, or administrative records to strengthen validation.
The fuzzy-set qualitative comparative analysis model treated high business financial inclusion as the outcome. Six causal conditions were incorporated into the configurational model: AI-enabled FinTech adoption, digital financial capability, perceived usefulness, familiarity and integration with artificial intelligence, digital financial skills, and digital financial self-efficacy. These conditions were selected because, jointly, they represent the technological adoption component and the capability-based mechanism through which SMEs can convert digital financial tools into effective inclusion. It should be noted that the use of a limited number of conditions was methodologically appropriate for the sample size, since it reduced excessive logical complexity in the truth table. Fuzzy calibration was performed using empirical percentiles, considering the distribution of Likert responses and the concentration of several variables in the upper range. The 25th percentile served as the threshold of full non-membership, the median as the crossover point, and the 75th percentile as the threshold of full membership. This strategy allowed fuzzy scores to reflect the actual empirical distribution of the sample while preserving a theoretically meaningful distinction between low, intermediate, and high membership in each condition. After calibration, all variables were transformed into fuzzy scores ranging from 0 to 1.
The analysis was conducted in two sequential stages. In the first stage, necessity analysis was performed to determine whether any individual condition, or its absence, constituted a requirement for the presence of high business financial inclusion. For each condition, consistency and coverage values were estimated, and the conventional consistency threshold of 0.90 was adopted as the criterion for identifying a necessary condition. The same procedure was replicated for the absence of the outcome, which made it possible to examine whether the configurations associated with low financial inclusion follow an asymmetric pattern. In the second stage, sufficiency analysis was carried out through truth-table construction and Boolean minimization. The truth table was built using a minimum frequency threshold of two cases per configuration, appropriate for the sample size, as well as a consistency threshold of 0.80 and a proportional reduction in inconsistency (PRI) threshold of 0.70. The intermediate solution was prioritized for substantive interpretation because it balances empirical evidence with theoretical plausibility. The complex and parsimonious solutions were also examined as complementary results to evaluate the stability of the configurational patterns. Model quality was assessed through consistency, raw coverage, unique coverage, and overall coverage. Finally, robustness tests were conducted by varying frequency and consistency thresholds to verify the stability of configurations under alternative analytical specifications.
4 Results
The final analytical sample consisted of 120 complete responses from owners, managers, or administrators of small and medium-sized enterprises. The fuzzy-set qualitative comparative analysis procedure was implemented after constructing composite scores for the main theoretical constructs and selected causal conditions. Given the high concentration of responses in the upper range of the Likert scale, the calibration strategy used empirical percentiles rather than fixed theoretical thresholds. Consequently, the 25th percentile operated as the anchor of full non-membership, the median as the crossover point, and the 75th percentile as the anchor of full membership. This decision was consistent with the distributional profile of the data, which showed relatively high means and negative skewness in the main constructs.
As reported in Table 1, the composite indicators showed adequate internal consistency. The three higher-order constructs presented Cronbach's alpha values above 0.94, with AI-enabled FinTech adoption at alpha = 0.961, digital financial capability at alpha = 0.954, and business financial inclusion at alpha = 0.941. The specific conditions retained for the configurational model also showed satisfactory reliability, with values ranging from alpha = 0.868 for familiarity and integration with artificial intelligence to alpha = 0.927 for perceived usefulness of AI-FinTech services. These results indicate that the constructed scales were internally coherent and appropriate for subsequent calibration based on set theory.
Table 1
| Construct/condition | Code | Mean | SD | alpha | Full non-memb. | Crossover | Full memb. |
|---|---|---|---|---|---|---|---|
| AI-enabled FinTech adoption | AFTIA | 3.647 | 0.821 | 0.961 | 3.319 | 3.778 | 4.069 |
| Digital financial capability | CFD | 3.683 | 0.825 | 0.954 | 3.250 | 3.813 | 4.188 |
| Perceived usefulness of AI-FinTech | UTI | 3.803 | 0.885 | 0.927 | 3.400 | 4.000 | 4.400 |
| Familiarity and integration with AI | FIA | 3.443 | 0.937 | 0.868 | 3.000 | 3.600 | 4.000 |
| Digital financial skills | HDF | 3.754 | 0.914 | 0.875 | 3.250 | 4.000 | 4.250 |
| Digital financial self-efficacy | AED | 3.702 | 0.891 | 0.872 | 3.250 | 4.000 | 4.063 |
| Business financial inclusion | IFE | 3.844 | 0.772 | 0.941 | 3.519 | 3.962 | 4.327 |
Descriptive statistics, internal consistency, and fuzzy calibration anchors.
As a supplementary diagnostic for common method variance, an unrotated single-factor solution was estimated using the fourteen available composite indicators. The first factor explained 73.39% of the total variance, which is above the conventional 50% reference value often used in Harman-type diagnostics. This result indicates that common method variance cannot be ruled out in the present self-report design. Accordingly, the results should be interpreted as exploratory configurational evidence and not as definitive causal proof. The high first-factor share may also reflect the theoretical proximity among technology adoption, digital capability, and financial inclusion, but it reinforces the need for future multi-source and longitudinal validation.
The descriptive profile also reveals that business financial inclusion achieved the highest composite mean among the main constructs, with M = 3.844 and SD = 0.772. Within its dimensions, quality and satisfaction showed the highest score, followed very closely by access to and use of digital financial services. In contrast, familiarity and integration with artificial intelligence obtained the lowest mean among the analysed conditions, with M = 3.443 and SD = 0.937. This pattern suggests that participating SMEs report relatively favourable levels of digital financial inclusion and perceived usefulness of FinTech services, although they exhibit a comparatively weaker level of familiarity with AI-based functionalities. Figure 1 illustrates this distributional structure, where the confidence intervals show that the most robust profiles are concentrated around financial inclusion, use, access, quality, and perceived usefulness, whereas familiarity with artificial intelligence remains the least developed component.
Figure 1
Necessity analysis was performed to determine whether any individual condition, or its absence, could be considered necessary for the presence of high business financial inclusion. As shown in Table 2, none of the analysed conditions reached the conventional consistency threshold of 0.90. Therefore, no individual causal condition can be interpreted as strictly necessary for high business financial inclusion. Nevertheless, digital financial capability exhibited the highest necessity consistency value, with 0.819 and coverage of 0.782, positioning it as the condition closest to necessity, although it remained below the accepted threshold. AI-enabled FinTech adoption also showed a relatively high value, with consistency of 0.793; it was followed by digital financial skills with 0.776 and familiarity and integration with artificial intelligence with 0.752.
Table 2
| Condition | Consistency high IFE | Coverage high IFE | Absence | Consistency low IFE | Coverage low IFE |
|---|---|---|---|---|---|
| AFTIA | 0.793 | 0.736 | ∼AFTIA | 0.741 | 0.798 |
| CFD | 0.819 | 0.782 | ∼CFD | 0.792 | 0.828 |
| UTI | 0.716 | 0.741 | ∼UTI | 0.772 | 0.750 |
| FIA | 0.752 | 0.690 | ∼FIA | 0.693 | 0.755 |
| HDF | 0.776 | 0.771 | ∼HDF | 0.790 | 0.795 |
| AED | 0.713 | 0.776 | ∼AED | 0.813 | 0.757 |
Necessity analysis for high and low business financial inclusion.
The conventional threshold for necessity is 0.90. The symbol “∼” denotes the absence of the condition. IFE, business financial inclusion.
The analysis of low business financial inclusion reinforces the asymmetric nature of the model. Although no absent condition reached the 0.90 threshold, the absence of digital financial self-efficacy exhibited the highest consistency for low inclusion, with 0.813. Similarly, the absence of digital financial capability, digital financial skills, and perceived usefulness presented consistency values close to 0.80. These results suggest that the lack of autonomy, capability, and digital financial skills is strongly associated with low financial inclusion, although none of these conditions operates individually as necessary. Figure 2 visually confirms this result, as all conditions remain to the left of the 0.90 threshold, although digital financial capability and AI-enabled FinTech adoption are the conditions closest to necessity for high business financial inclusion.
Figure 2
The truth table was constructed using the six causal conditions and high business financial inclusion as the outcome. The model applied a frequency cut-off of two cases, a consistency cut-off of 0.80, and a PRI cut-off of 0.70. Under these criteria, three empirical configurations were classified as sufficient for the outcome, with no contradictory configurations. The most populated configuration corresponded to full presence, composed of high AI-FinTech adoption, high digital financial capability, high perceived usefulness, high familiarity with artificial intelligence, high digital financial skills, and high digital financial self-efficacy. This configuration grouped 38 cases and recorded consistency of 0.862 and PRI of 0.803. Two additional configurations also exceeded the sufficiency threshold: one characterized by the presence of all core conditions except familiarity with artificial intelligence; the other by the presence of all core conditions except digital financial self-efficacy. Figure 3 graphically presents these three configurations.
Figure 3
The intermediate minimization solution offered a theoretically more meaningful representation of the sufficient pathways. As shown in Table 3, the solution identified two alternative configurations leading to high business financial inclusion. The first pathway combines high AI-FinTech adoption, high digital financial capability, high perceived usefulness, high familiarity and integration with artificial intelligence, and high digital financial skills. This configuration reached consistency of 0.870, PRI of 0.820, and raw coverage of 0.543. The second pathway combines high AI-FinTech adoption, high digital financial capability, high perceived usefulness, high digital financial skills, and high digital financial self-efficacy, with consistency of 0.866, PRI of 0.807, and raw coverage of 0.523. The overall solution reached consistency of 0.873, PRI of 0.823, and coverage of 0.582, indicating that the identified causal recipes explain a substantial portion of membership in high business financial inclusion.
Table 3
| Pathway | Sufficient configuration | Consistency | PRI | Raw cov. | Unique cov. |
|---|---|---|---|---|---|
| P1 | AFTIA * CFD * UTI * FIA * HDF | 0.870 | 0.820 | 0.543 | 0.059 |
| P2 | AFTIA * CFD * UTI * HDF * AED | 0.866 | 0.807 | 0.523 | 0.040 |
| Overall solution | P1 + P2 | 0.873 | 0.823 | 0.582 | - |
Intermediate solution for high business financial inclusion.
The symbol “*” denotes logical conjunction (AND). AFTIA, AI-enabled FinTech adoption; CFD, digital financial capability; UTI, perceived usefulness; FIA, familiarity and integration with artificial intelligence; HDF, digital financial skills; AED, digital financial self-efficacy. PRI, proportional reduction in inconsistency; Cov., coverage.
The two pathways show that high business financial inclusion among SMEs does not arise from isolated technology adoption. Rather, the outcome emerges from a combined structure in which adoption is accompanied by perceived usefulness, digital financial capability, and operational digital skills. The repeated presence of AFTIA, CFD, UTI, and HDF in both pathways indicates that these conditions constitute the central configurational core of the model. Familiarity with artificial intelligence and digital financial self-efficacy appear as alternative reinforcing mechanisms: in one pathway, familiarity with AI-based functionalities strengthens the conversion of FinTech adoption into inclusion; in the other pathway, self-efficacy compensates by enabling SMEs to operate digital financial services with autonomy and confidence.
The analysis of low business financial inclusion was expanded to evaluate the causal-asymmetry claim more transparently. As shown in Table 4, the truth table for low inclusion identified seven empirical sufficient rows that exceeded the inclusion threshold, indicating that low inclusion is not limited to a single empirical profile. These rows include a cumulative-deficit profile, where the six conditions are absent, as well as mixed-deficit profiles in which one or two conditions appear present but are not enough to compensate for the absence of digital financial capability, AI-FinTech adoption, AI familiarity, skills, or self-efficacy. The most populated empirical row corresponded to the simultaneous absence of the six conditions, grouping 27 cases and reaching consistency of 0.915. However, additional rows show that low inclusion can also occur when perceived usefulness, digital financial skills, or self-efficacy are present in isolation, because these isolated strengths do not create inclusion when the broader capability-conversion architecture is missing.
Table 4
| Row | Empirical configuration for low IFE | n | incl. | Interpretation |
|---|---|---|---|---|
| 12 | ∼AFTIA * ∼CFD * UTI * ∼FIA * HDF * AED | 2 | 0.945 | Mixed deficit |
| 5 | ∼AFTIA * ∼CFD * ∼UTI * FIA * ∼HDF * ∼AED | 2 | 0.932 | Mixed deficit |
| 1 | ∼AFTIA * ∼CFD * ∼UTI * ∼FIA * ∼HDF * ∼AED | 27 | 0.915 | Cumulative deficit |
| 41 | AFTIA * ∼CFD * UTI * ∼FIA * ∼HDF * ∼AED | 2 | 0.892 | Capability deficit despite adoption |
| 3 | ∼AFTIA * ∼CFD * ∼UTI * ∼FIA * HDF * ∼AED | 6 | 0.875 | Skills isolated from capability |
| 9 | ∼AFTIA * ∼CFD * UTI * ∼FIA * ∼HDF * ∼AED | 7 | 0.860 | Usefulness isolated from adoption/capability |
| 2 | ∼AFTIA * ∼CFD * ∼UTI * ∼FIA * ∼HDF * AED | 3 | 0.847 | Self-efficacy isolated from adoption/capability |
Empirical truth-table configurations sufficient for low business financial inclusion.
The symbol “∼” denotes absence of the condition. The table reports empirical truth-table rows classified as sufficient for low business financial inclusion before Boolean minimization.
As shown in Table 5, the minimized intermediate solution confirms that low inclusion is organized around two reduced pathways: the joint absence of AI-FinTech adoption and digital financial capability, and the joint absence of digital financial capability and AI familiarity. Both pathways centre on the absence of digital financial capability, reinforcing its role as the main conversion mechanism in the model. Because the high-inclusion solution requires the presence of adoption, capability, usefulness, and skills, while the low-inclusion solution is explained by capability deficits combined with absent adoption or absent AI familiarity, the evidence supports causal asymmetry rather than a simple inverse mirror of the high-inclusion recipes.
Table 5
| Pathway | Sufficient configuration | Consistency | PRI | Raw cov. | Unique cov. |
|---|---|---|---|---|---|
| L1 | ∼AFTIA * ∼CFD | 0.866 | 0.840 | 0.691 | 0.066 |
| L2 | ∼CFD * ∼FIA | 0.850 | 0.825 | 0.657 | 0.032 |
| Overall solution | L1 + L2 | 0.852 | 0.824 | 0.723 | n/a |
Intermediate solution for low business financial inclusion.
AFTIA, AI-enabled FinTech adoption; CFD, digital financial capability; FIA, familiarity and integration with artificial intelligence; PRI, proportional reduction in inconsistency; Cov., coverage.
A robustness assessment was carried out by varying the inclusion and frequency thresholds. As reported in Table 6, the solution remained stable across alternative specifications. When the inclusion threshold was set at 0.80 or 0.85, three sufficient configurations were retained for all frequency thresholds. When the inclusion threshold was increased to 0.90, two sufficient configurations were preserved. It is worth emphasizing that no contradictory configurations were observed in any robustness scenario. These results support the internal consistency of the configurational findings and suggest that the main solution is not an artefact of a single threshold decision.
Table 6
| Inclusion threshold | Frequency threshold | Sufficient configurations | Contrad. | Logical remainders |
|---|---|---|---|---|
| 0.80 | 1 | 3 | 0 | 37 |
| 0.85 | 1 | 3 | 0 | 37 |
| 0.90 | 1 | 2 | 0 | 37 |
| 0.80 | 2 | 3 | 0 | 50 |
| 0.85 | 2 | 3 | 0 | 50 |
| 0.90 | 2 | 2 | 0 | 50 |
| 0.80 | 3 | 3 | 0 | 56 |
| 0.85 | 3 | 3 | 0 | 56 |
| 0.90 | 3 | 2 | 0 | 56 |
Robustness assessment using alternative truth-table thresholds.
The results correspond to the intermediate solution under nine alternative specifications. The stability of the solution across thresholds supports the robustness of the configurational model.
The results indicate that high business financial inclusion among SMEs is configurational rather than unidimensional in nature. The findings support the idea that AI-enabled FinTech adoption contributes to financial inclusion only when embedded in a broader capability structure. Specifically, the most stable explanation combines adoption, perceived usefulness, digital financial capability, and digital financial skills, whereas familiarity with artificial intelligence and digital financial self-efficacy operate as alternative mechanisms that complete the pathway towards high inclusion. Consequently, the configurational evidence complements the mediation logic of the study by showing that digital financial capability is not merely an accessory intervening variable, but a central condition within the causal architecture that enables SMEs to transform the use of AI-enabled FinTech into effective financial inclusion.
5 Discussion
This study advances a capability-conversion interpretation of AI-enabled FinTech adoption and business financial inclusion. The two high-inclusion pathways show that adoption is theoretically consequential only when it is embedded in a broader architecture of perceived usefulness, digital financial capability, and digital financial skills. This finding refines Technology Acceptance Theory by showing that perceived usefulness and adoption explain the opening of technological opportunities, but not the full conversion of those opportunities into inclusion. In other words, acceptance is a necessary part of the story, but it is not the final mechanism that produces business financial inclusion.
Regarding AI-enabled FinTech adoption, the findings indicate that artificial intelligence matters because it changes the type of financial opportunity available to SMEs. Algorithmic credit assessment, predictive financial analytics, automated recommendations, fraud detection, and AI-assisted decision interfaces may reduce information frictions and personalize services, but they also require firms to interpret algorithmic outputs and act on them. Thus, the results support the technology acceptance literature while extending it: AI-enabled FinTech adoption should be understood less as a simple use decision and more as a gateway into algorithmically mediated financial participation. The presence of familiarity with artificial intelligence in one pathway suggests that firms can convert adoption into inclusion more effectively when they understand and trust AI-based functionalities, rather than treating them as opaque digital tools.
Regarding digital financial capability, the findings strengthen the capability approach by positioning capability as the conversion mechanism between technological access and achieved inclusion. The repeated presence of digital financial capability and digital financial skills in the high-inclusion pathways indicates that firms require more than exposure to platforms: they need the ability to navigate interfaces, evaluate fees and conditions, monitor transactions, manage authentication, and use digital information for financial decisions. This interpretation extends prior work on digital financial literacy and financial capability by showing that capability operates configurationally; it becomes effective when combined with adoption and perceived usefulness, and not as an isolated individual attribute.
The asymmetry analysis also extends financial inclusion theory. Low business financial inclusion was not explained by the simple absence of the high-inclusion solution. Instead, the low-inclusion solution was organized around the absence of digital financial capability combined with either the absence of AI-FinTech adoption or the absence of AI familiarity. This means that exclusion should be understood as a cumulative and configurational deficit, not as a single missing input. From a theoretical perspective, this finding challenges diffusion-centred assumptions that expanding digital financial technologies will automatically reduce exclusion. From a policy perspective, it implies that interventions should not only promote adoption but also build the capabilities that allow firms to interpret, trust, and use AI-enabled financial tools effectively.
Overall, the study transforms empirical confirmation into theory extension in three ways. First, it reframes AI-enabled FinTech adoption as an opportunity structure rather than a direct cause of inclusion. Second, it specifies digital financial capability as the mechanism that converts opportunity into financial participation and value creation. Third, it demonstrates that this conversion occurs through equifinal and asymmetric configurations rather than through a single linear pathway. These arguments refine Technology Acceptance Theory by identifying its boundary condition, extend Capability Theory by applying conversion logic to AI-enabled financial services, and deepen Financial Inclusion Theory by treating inclusion as an achieved configurational outcome.
6 Conclusions
This study advances knowledge by demonstrating that business financial inclusion is not a direct consequence of AI-enabled FinTech adoption. Rather, it is the outcome of a capability-conversion process in which firms transform algorithmically enabled financial opportunities into effective access, use, quality, and satisfaction with formal financial services. The high-inclusion solution identified two equifinal pathways with overall coverage of 0.582 and consistency of 0.873. In both pathways, AI-enabled FinTech adoption, digital financial capability, perceived usefulness, and digital financial skills formed the common causal core, while AI familiarity and digital financial self-efficacy operated as alternative reinforcing mechanisms. This confirms that the key theoretical issue is not adoption alone, but the configuration that enables SMEs to convert adoption into inclusion.
The first knowledge contribution concerns Technology Acceptance Theory. The findings support the relevance of adoption and perceived usefulness, but also show that acceptance-based constructs are insufficient to explain achieved financial inclusion. In AI-enabled FinTech environments, adoption creates access to algorithmic services such as automated credit assessment, predictive analytics, intelligent recommendations, and assisted decision interfaces. However, these opportunities produce inclusion only when firms also possess the capability and skills required to understand and use them. Therefore, the study identifies the boundary of adoption-centred explanations and shows why similar levels of AI-FinTech adoption may generate different inclusion outcomes.
The second contribution concerns Capability Theory. Digital financial capability was the condition closest to necessity and appeared in both high-inclusion pathways. This supports the conclusion that capability is not an accessory antecedent, but the core conversion mechanism through which AI-enabled FinTech adoption becomes meaningful financial participation. The presence of digital financial skills in both pathways further indicates that conversion depends on operational competence, not only on knowledge or positive attitudes. The alternative role of AI familiarity and self-efficacy shows that SMEs can achieve inclusion through different capability profiles, provided that the central conversion architecture is present.
The third contribution concerns Financial Inclusion Theory and configurational reasoning. The analysis of low inclusion showed that exclusion is not the symmetrical inverse of inclusion. The intermediate solution for low inclusion was explained by the absence of digital financial capability combined with either the absence of AI-FinTech adoption or the absence of AI familiarity, with overall coverage of 0.723 and consistency of 0.852. This supports the conclusion that business financial exclusion is generated by cumulative capability and adoption deficits, rather than by a single missing condition. Consequently, policies aimed at inclusion should combine technology diffusion with training in digital financial skills, interpretive capability for AI-based recommendations, and self-efficacy for autonomous use of digital financial services.
The study presents limitations that should be considered when interpreting its contributions. First, the cross-sectional design restricts the possibility of inferring temporal dynamics in the pathway from technology adoption to financial inclusion. Second, the non-probability purposive sample of 120 SMEs is appropriate for exploratory fsQCA but does not allow statistical representativeness or direct generalization to all firms in northern Peru or Latin America. Third, because recruitment did not generate a closed denominator of invited eligible firms, a conventional response rate could not be calculated. Fourth, all measurements came from self-report instruments administered to one owner, manager, or administrator per firm, which creates the risk of social desirability bias and common method variance; the single-factor diagnostic indicated that this risk cannot be ruled out. Fifth, no marker variable or multi-source validation was included, and no administrative or banking records were available to corroborate the reported practices. Sixth, fuzzy calibration based on empirical percentiles, despite its adequacy for the observed distribution, constitutes a methodological decision that may dialogue differently with alternative theoretical calibrations.
Future research should advance this knowledge agenda in five directions. First, longitudinal or rotating-panel designs should examine how the capability-conversion pathways evolve as AI-enabled financial services mature. Second, probability-based and regionally expanded samples should test whether the identified configurations hold across other Peruvian and Latin American SME ecosystems. Third, mixed analytical designs should combine fsQCA with structural equation modelling to triangulate average effects with sufficient configurations. Fourth, future studies should include multi-respondent data, marker variables, and administrative or banking records to reduce common method bias. Fifth, the technological domain should be expanded toward generative AI, open banking, blockchain-based alternative finance, and embedded finance to evaluate whether new algorithmic features create additional conversion requirements for business financial inclusion.
Statements
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Ethics statement
Ethical approval was not required for the study involving humans in accordance with local legislation and institutional requirements. Written informed consent was obtained from all participants before questionnaire completion. Participation was voluntary and anonymous, and no personally identifiable information was collected.
Author contributions
LC: Writing – original draft, Writing – review & editing. AH: Writing – original draft; Writing – review & editing. JT: Writing – original draft, Writing – review & editing. DC: Writing – original draft, Writing – review & editing. NB: Writing – original draft, Writing – review & editing. OC: Writing – original draft, Writing – review & editing. TF: Writing – original draft, Writing – review & editing. JI: Writing – original draft, Writing – review & editing.
Funding
The author(s) declared that financial support was not received for this work and/or its publication.
Acknowledgments
The authors acknowledge the collaboration of the owners, managers, and administrators of small and medium-sized enterprises in northern Peru who voluntarily participated in the study. The authors also acknowledge the academic and technical support received during the preparation, organization, and review of the manuscript.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1
AkhtarM.GhafoorK. A.MaqsoodH.MurtazaK. (2026). Fintech payment services adoption among women: the impact of personality, financial services knowledge and artificial intelligence. Int. J. Account. Inf. Manag.34 (1), 193–219. 10.1108/IJAIM-09-2024-0355
2
AkhtarM.SalmanA.Abdul GhafoorK.KamranM. (2024). Artificial intelligence, financial services knowledge, government support, and user innovativeness: exploring the moderated-mediated path to fintech adoption. Heliyon10 (20), e39521. 10.1016/j.heliyon.2024.e39521
3
AlkadiR. S.AbedS. S. (2025). AI In banking: what drives generation Z to adopt AI-enabled voice assistants in Saudi Arabia?Int. J. Financ. Stud.13 (1), 36. 10.3390/ijfs13010036
4
AmnasM. B.SelvamM.ParayitamS. (2024). Fintech and financial inclusion: exploring the mediating role of digital financial literacy and the moderating influence of perceived regulatory support. J. Risk Financ. Manag.17 (3), 108. 10.3390/jrfm17030108
5
AnhN. Q. (2026). An integrated model of financial socialization, technology, and financial capability in predicting financial well-being. PLoS One21 (3), e0340002. 10.1371/journal.pone.0340002
6
AppiahT.AgblewornuV. V. (2025). The interplay of perceived benefit, perceived risk, and trust in FinTech adoption: insights from Sub-Saharan Africa. Heliyon11 (4), e41992. 10.1016/j.heliyon.2025.e41992
7
AppiahT.AssefuahB.AgblewornuV. V.ForsonN. M. A.DonkohG. (2025). Adoption of electronic payment platforms by fast moving consumer goods (FMCG) businesses in Ghana: PLS-SEM and fsQCA approach. SN Bus. Econ.5 (8), 156. 10.1007/s43546-025-00820-z
8
AsandimitraN.KautsarA.KusumaningrumT. M.ArifahJ. (2023). Undermining shoestring budget: financial capability determinants of millennial generation. Int. J. Sustain. Dev. Plan.18 (4), 1287–1296. 10.18280/ijsdp.180417
9
AwE. C. X.ZhaT.ChuahS. H. W. (2023). My new financial companion! Non-linear understanding of robo-advisory service acceptance. Serv. Ind. J.43 (3-4), 185–212. 10.1080/02642069.2022.2161528
10
Banco Central de Reserva del Perú (2025). Reporte de estabilidad financiera, mayo 2025. Banco Central de Reserva del Perú. Available online at:https://www.bcrp.gob.pe/publicaciones/reporte-de-estabilidad-financiera/ref-mayo-2025.html(Accessed April 15, 2026).
11
BanduraA. (1997). Self-efficacy: The Exercise of Control. New York, NY: W. H. Freeman.
12
BayakhmetovaA.RudenkoL.KrylovaL.SuleimenovaB. (2025). Artificial intelligence in financial behavior: bibliometric ideas and new opportunities. J. Risk Financ. Manag.18 (3), 159. 10.3390/jrfm18030159
13
BojuwonM.OlaleyeB. R.OjebodeA. A. (2023). Financial inclusion and financial condition: the mediating effect of financial self-efficacy and financial literacy. Vis. J. Bus. Perspect.27 (5), 642–654. 10.1177/09722629231166200
14
DavisF. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS. Q.13 (3), 319–340. 10.2307/249008
15
DuarteL. L.BaptistaG. (2026). Artificial intelligence: understanding the determinants of adoption and intention to recommend the technology in the banking sector. Int. J. Bank Mark.44 (3), 492–517. 10.1108/IJBM-06-2024-0372
16
FanL. (2021). A conceptual framework of financial advice-seeking and short- and long-term financial behaviors: an age comparison. J. Fam. Econ. Issues.42 (1), 90–112. 10.1007/s10834-020-09727-3
17
FanL.ZengY. (2025). Financial education mandates and financial wellbeing of young adults: examining financial capability factors as mediators. Int. J. Consum. Stud.49 (2), e70058. 10.1111/ijcs.70058
18
Finnovista (2025). Finnovista Fintech Radar Perú 2024. Finnovista, Mastercard, and Galileo. Available online at:https://www.finnosummit.com/radar/finnovista-fintech-radar-peru-2024/(Accessed April 15, 2026).
19
FissP. C. (2011). Building better causal theories: a fuzzy set approach to typologies in organization research. Acad. Manage. J.54 (2), 393–420. 10.5465/amj.2011.60263120
20
GrableJ.KwakE. J.ArchuletaK. (2023). Distrust of banks among the unbanked and banked. Int. J. Bank Mark.41 (5), 1107–1124. 10.1108/IJBM-10-2022-0441
21
GuoY.WakedH. N. (2026). Trust-mediated adoption of AI robo-advisors in inland China: an extended UTAUT perspective. J. Financ. Serv. Mark.31 (11), 1–1310.1057/s41264-025-00339-2
22
HamurcuÇHazarA.BabuşcuŞ (2025). A portfolio-focused behavioral model linking financial literacy and risk tolerance: evidence from mutual fund investors. Borsa Istanb. Rev.25 (Suppl. 2), 119–127. 10.1016/j.bir.2025.09.001
23
HentzenJ. K.HoffmannA.DolanR.PalaE. (2022). Artificial intelligence in customer-facing financial services: a systematic literature review and agenda for future research. Int. J. Bank Mark.40 (6), 1299–1336. 10.1108/IJBM-09-2021-0417
24
HjelkremL. O.de LangeP. E.NessetE. (2023). Explaining deep learning models for credit scoring with SHAP: a case study using open banking data. J. Risk Financ. Manag.16 (4), 221. 10.3390/jrfm16040221
25
IgamoA. M.RachmatR. A.SiregarM. I.GaribaM. I.CheronoV.WahyuniA. S.et al (2024). Factors influencing Fintech adoption for women in the post-COVID-19 pandemic. J. Open Innov.: Technol., Mark. Complex.10 (2), 100236. 10.1016/j.joitmc.2024.100236
26
Inter-American Development Bank and Finnovista (2024). Fintech in Latin America and the Caribbean: A consolidated ecosystem with potential to contribute to regional financial inclusion. Inter-American Development Bank.
27
JenaR. K. (2025). Factors influencing the adoption of FinTech for the enhancement of financial inclusion in rural India using a mixed methods approach. J. Risk Financ. Manag.18 (3), 150. 10.3390/jrfm18030150
28
KumarP.PillaiR.KumarN.TabashM. I. (2023). The interplay of skills, digital financial literacy, capability, and autonomy in financial decision making and well-being. Borsa Istanb. Rev.23 (1), 169–183. 10.1016/j.bir.2022.09.012
29
KurniasariF.LestariE. D.MuslimN. A.Mohd AliM. A. (2026). From knowledge to prosperity: financial literacy driving women entrepreneur led-small medium enterprises through technology innovation. Soc. Sci. Humanit. Open13, 102443. 10.1016/j.ssaho.2026.102443
30
LeeC. T.PanL. Y. (2023). Smile to pay: predicting continuous usage intention toward contactless payment services in the post-COVID-19 era. Int. J. Bank Mark.41 (2), 312–332. 10.1108/IJBM-03-2022-0130
31
Ministerio de la Producción del Perú (2025). Indicadores de desempeño MYPE 2024. Gobierno del Perú. Available online at:https://www.producempresarial.pe/(Accessed April 15, 2026).
32
MishraD.AgarwalN.SharahileyS.KandpalV. (2024a). Digital financial literacy and its impact on financial decision-making of women: evidence from India. J. Risk Financ. Manag.17 (10), 468. 10.3390/jrfm17100468
33
MishraD.KandpalV.AgarwalN.SrivastavaB. (2024b). Financial inclusion and its ripple effects on socio-economic development: a comprehensive review. J. Risk Financ. Manag.17 (3), 105. 10.3390/jrfm17030105
34
MittalK.GuptaS. K.GuptaK.KumarS.KumarA. (2025). The digital shift in finance: examining FinTech adoption through PLS-SEM and fsQCA analysis. Asia Pac. J. Inf. Syst.35 (3), 681–704. 10.14329/apjis.2025.35.3.681
35
MladenovićD.MonashevM.JirásekM.BruniR. (2025). When online review is not enough? Adoption of cryptocurrencies through the lenses of NCA and fsQCA. Comput. Human. Behav.168, 108764. 10.1016/j.chb.2025.108764
36
MuatS.MahdzanN. S.SukorM. E. A.FachrurroziF.SariN. (2025). The role of financial inclusion, financial literacy and digital payment adoption in Indonesian millennials’ financial well-being. Int. J. Bank Mark.43 (9), 1938–1968. 10.1108/IJBM-01-2024-0022
37
OsakweC. N.OgunmokunO. A.ElgammalI.KwartengM. A. (2025). Individuals’ attitudes and their adoption intentions of central bank digital currency: combining theories and analytical techniques. Technol. Forecast. Soc. Change.210, 123857. 10.1016/j.techfore.2024.123857
38
PappasI. O.WoodsideA. G. (2021). Fuzzy-set qualitative comparative analysis (fsQCA): guidelines for research practice in information systems and marketing. Int. J. Inf. Manage.58, 102310. 10.1016/j.ijinfomgt.2021.102310
39
ParkS.YoonS. (2025). Cross-national findings of factors affecting the acceptance of AI-based sustainable FinTech. Sustainability17 (1), 49. 10.3390/su17010049
40
PeterS.GeethaE.GuptaA. (2024). Navigating the digital financial landscape: unraveling the impact of financial behavior traits on women-owned enterprises in the new normal. Cogent Bus. Manag.11 (1), 2296570. 10.1080/23311975.2023.2296570
41
PratamaR.WijayantiR.IndrawatiN. K. (2026). Behavior-first finance: reframing digital literacy and self-efficacy as drivers of MSME well-being. Multidiscip. Sci. J.8 (5), e2026430. 10.31893/multiscience.2026430
42
RaginC. C. (2008). Redesigning Social Inquiry: Fuzzy Sets and Beyond. Chicago, IL: University of Chicago Press.
43
RohT.ParkB. I.XiaoS. (2023). Adoption of AI-enabled robo-advisors in FinTech: simultaneous employment of UTAUT and the theory of reasoned action. J. Electron. Commer. Res.24 (1), 29–47.
44
SabriM. F.TrinugrohoI.MagliA. S.LawS. H.WahyonoB.RidhwanM. M.et al (2026). Bridging digital capabilities and financial resilience: a comparative study of FinTech, AI literacy, and financial inclusion among households in Malaysia and Indonesia. J. Open Innov.: Technol., Mark. Complex.12 (2), 100759. 10.1016/j.joitmc.2026.100759
45
SajidM.MushtaqR.MurtazaG.YahiaouiD.PereiraV. (2024). Financial literacy, confidence and well-being: the mediating role of financial behavior. J. Bus. Res.182, 114791. 10.1016/j.jbusres.2024.114791
46
SenA. (1999). Development as Freedom. Oxford: Oxford University Press.
47
SetiawanB.RaniS.EmildaE.ArifinF.UtamiD. (2026). Understanding FinTech adoption drivers for digital financial sustainability in urban and rural MSMEs. Risks14 (4), 77. 10.3390/risks14040077
48
SetiawanB.TrianaD.Al AzizahU. S.WahyuniA. S.VictorV.NathanR. J.et al (2025). Financial technology (Fintech) innovation and financial inclusion: comparative study of urban and rural consumers post-COVID-19 pandemic. J. Innov. Entrep.14 (1), 1–28. 10.1186/s13731-024-00452-x
49
SinghJ.BatraG. S.KaurS. (2025). Forecasting inclusive futures: fintech, capability expansion, and livelihood pathways in urban India. Technol. Forecast. Soc. Change.215, 124369. 10.1016/j.techfore.2025.124369
50
SunS.ChenY. C.AnsongD.HuangJ.SherradenM. S. (2022). Household financial capability and economic hardship: an empirical examination of the financial capability framework. J. Fam. Econ. Issues.43 (4), 716–729. 10.1007/s10834-022-09816-5
51
Superintendencia de Banca, Seguros y AFP (2025). Informe de estabilidad del sistema financiero, mayo 2025. Superintendencia de Banca, Seguros y AFP. Available online at:https://www.sbs.gob.pe/Portals/0/IESF-2025-1A.pdf(Accessed April 15, 2026).
52
TandilinoC.PontohG. T.DarmawatiD.IndrijawatiA. (2025). Digital financial inclusion as a mediator of digital financial literacy and government support in MSME performance. Int. J. Financ. Stud.13 (4), 199. 10.3390/ijfs13040199
53
TettehF. K.KwatengK. O. (2026). The pathways from digital financial literacy to sustained engagement with mobile financial services: a technology continuance theory perspective. J. Financ. Serv. Mark.31 (7), 1–22. 10.1057/s41264-025-00335-6
54
United Nations (2015). Transforming our World: The 2030 Agenda for Sustainable Development. New York, NY: United Nations.
55
VenkateshV.MorrisM. G.DavisG. B.DavisF. D. (2003). User acceptance of information technology: toward a unified view. MIS. Q.27 (3), 425–478. 10.2307/30036540
56
VenkateshV.ThongJ. Y. L.XuX. (2012). Consumer acceptance and use of information technology: extending the unified theory of acceptance and use of technology. MIS. Q.36 (1), 157–178. 10.2307/41410412
57
World Bank (2025). The Global Findex Database 2025: Connectivity and financial inclusion in the digital economy. World Bank Group. Available online at:https://www.worldbank.org/en/publication/globalfindex(Accessed April 15, 2026).
58
XiaoJ. J.MengK. (2024). Financial capability and financial anxiety: comparison before and during the COVID-19 pandemic. Int. J. Bank Mark.42 (3), 565–587. 10.1108/IJBM-03-2023-0140
59
YangX.YangJ.HouY.LiS.SunS. (2023). Gamification of mobile wallet as an unconventional innovation for promoting FinTech: an fsQCA approach. J. Bus. Res.155, 113406. 10.1016/j.jbusres.2022.113406
60
ZaimovicA.OmanovicA.Nuhic MeskovicM.Arnaut BeriloA.ZaimovicT.DedovicL.et al (2025). The nexus between digital financial knowledge and financial inclusion: digital financial attitudes and behaviours as mediators. Int. J. Bank Mark.43 (2), 388–423. 10.1108/IJBM-01-2024-0053
61
ZhangY.FanL. (2024). The nexus of financial education, literacy and mobile fintech: unraveling pathways to financial well-being. Int. J. Bank Mark.42 (7), 1789–1817. 10.1108/IJBM-09-2023-0531
Summary
Keywords
AI-enabled FinTech adoption, business financial inclusion, digital financial capability, digital financial skills, financial technology, fuzzy-set qualitative comparative analysis, northern Peru, small and medium-sized enterprises
Citation
Correa Rojas L, Haro Sarango A, Tejada Carrera JA, Castro Vargas DJ, Barboza Chuquilín NA, Cabrera Cabrera OH, Fernández Miranda TK and Izquierdo Espinoza JR (2026) Configurational pathways from AI-enabled FinTech adoption to business financial inclusion in SMEs: the role of digital financial capability in northern Peru. Front. Hum. Dyn. 8:1909382. doi: 10.3389/fhumd.2026.1909382
Received
15 June 2026
Revised
29 June 2026
Accepted
30 June 2026
Published
21 July 2026
Volume
8 - 2026
Edited by
Serhan S. Alshammari, University of Hail, Saudi Arabia
Reviewed by
Detak Prapanca, Universitas Muhammadiyah Sidoarjo, Indonesia
Yonas Ferdinand Riwu, University of Nusa Cendana, Indonesia
Updates
Copyright
© 2026 Correa Rojas, Haro Sarango, Tejada Carrera, Castro Vargas, Barboza Chuquilín, Cabrera Cabrera, Fernández Miranda and Izquierdo Espinoza.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Liliana Correa Rojas lcorrearo@unitru.edu.pe
ORCID Liliana Correa Rojas orcid.org/0000-0002-9505-8082 Alexander Haro Sarango orcid.org/0000-0001-7398-2760 Jorge Alejandro Tejada Carrera orcid.org/0000-0002-5255-6487 Daniel Jesús Castro Vargas orcid.org/0000-0003-0618-6013 Nixon Arnaldo Barboza Chuquilín orcid.org/0000-0002-9678-4369 Olegario Heiner Cabrera Cabrera orcid.org/0000-0002-2500-0627 Tattiana Katerine Fernández Miranda orcid.org/0000-0001-7504-3724 Julio Roberto Izquierdo Espinoza orcid.org/0000-0001-6827-273X
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.