Driving Sustainable AI Implementation in Business: The Integrated Role of Economic Value Perception, Managerial Attitudes, and Behavioral Intentions

Abstract

This research examines the relationships among economic value assessment, managerial perspectives, adoption willingness, and long-term AI utilization among organizational leaders in Peru. Using a quantitative cross-sectional design with covariance-based statistical analysis, data were collected from 390 management personnel (58.72% male, 41.28% female) representing diverse enterprises in northern Peru. Four hypotheses were evaluated concerning the effects of price–value on intention to use, attitude toward AI on intention to use, price–value on attitude toward AI, and intention to use on sustainable AI implementation. Results from partial-least-squares structural equation modeling showed significant direct effects of price–value on both intentions to use and attitudes toward AI, with attitudes also significantly influencing intention to use. The model exhibited exceptional explanatory power: price–value and attitudes explained 89.2% of the variance in intention to use, while intention to use accounted for 85.1% of the variance in sustainable AI use. These findings indicate that both economic considerations and psychological factors are critical for advancing sustainable AI adoption among business managers. Consequently, integrating price–value analyses and attitude-development components into AI implementation strategies is supported to enhance technology adoption success in business contexts.

1. Introduction

The examination of how cost–benefit evaluations, executive attitudes, and implementation readiness influence organizational leaders regarding artificial intelligence has become central to contemporary sustainable technology adoption research. Contemporary investigations reveal that cost assessments and managerial viewpoints demonstrate positive correlations with enduring technology utilization and deployment effectiveness [1,2]—factors essential for executives navigating intricate digital transformation challenges. Quantitative findings suggest that financial evaluations mediate associations between technology acceptance and continued deployment [3], while institutional perspectives affect how implementation beliefs influence lasting usage behaviors [4]. Furthermore, ref. [5] documented that economic value can strengthen sustainable utilization through flexible management approaches, and [6] indicated that monetary considerations affect both perspectives and deployment outcomes.

Existing scholarly research exposes notable conceptual and procedural limitations that constrain comprehension of AI adoption processes, especially concerning the synthesis of financial and cognitive elements in enduring implementation frameworks. While earlier studies have established separate linkages between financial assessments and technology acceptance [7], these investigations encounter three primary limitations constituting the central theoretical gap this research examines. Initially, academic research lacks integrated frameworks that concurrently analyze how financial considerations (price–value) and cognitive factors (attitudes) combine to affect behavioral intentions and sustained usage results. Most investigations study these constructs separately, missing the intricate relationships that define actual executive decision-making processes.

Subsequently, the temporal aspect of enduring AI utilization remains conceptually underdeveloped. Refs. [8,9] recognize that economic value considerations forecast implementation success, yet their framework approaches sustainability as a fixed result rather than a fluid process shaped by changing executive perceptions and perspectives. This represents a fundamental theoretical limitation, as sustainable technology implementation requires understanding how initial adoption intentions translate into long-term usage patterns through mediating psychological mechanisms.

Finally, organizational management scholarship exhibits conflicting theoretical models when investigating AI implementation. Refs. [10,11] present divergent results concerning financial-adoption associations, indicating that current theoretical frameworks may be insufficiently defined for modern AI deployment contexts. The absence of theoretically integrated frameworks that account for both economic rationality and psychological factors creates a significant knowledge gap that limits our ability to predict and explain sustainable AI adoption among business managers.

These conceptual constraints collectively indicate the necessity for a unified model that connects financial and cognitive perspectives while addressing the evolving nature of sustainable technology deployment. This research fills this void by developing and evaluating an inclusive framework that investigates how economic value perceptions affect attitudes, which subsequently form behavioral intentions and finally establish sustainable AI usage patterns among business executives.

The psychological dimension of sustainable technology implementation has developed as a fundamental framework for comprehending cognitive factors that enhance executives’ effectiveness and sustained utilization [12]. This perspective acknowledges that organizational managers must cultivate not only technical skills but also suitable perspectives and value assessments to tackle complex deployment challenges [13]. Investigations in this domain indicate that economic value assessments and favorable perspectives are especially crucial for executives, as they frequently encounter implementation apprehension, deployment obstacles, and institutional pressures of promoting sustainable AI utilization [14,15,16].

The main goal of this investigation is to examine associations between economic value, perspectives, and adoption willingness within sustainable AI implementation among organizational managers. This research develops four core hypotheses: the effect of economic value on adoption willingness, the impact of perspectives on adoption willingness, the influence of economic value on perspectives, and the role of adoption willingness on sustainable AI deployment. Its innovation resides in emphasizing organizational executives and the concurrent examination of both financial and cognitive factors, elements previously unexplored in sustainable AI implementation scholarship.

The conceptual rationale for this investigation rests in its contribution to comprehending how economic value perceptions and perspectives may affect sustainable AI utilization in organizational contexts, building upon [17] research on the role of financial beliefs in technology implementation. From an applied standpoint, the results could guide the creation of more efficient strategies to foster sustainable AI utilization among organizational executives, consistent with [18,19]. As recommended by [20,21] incorporating both financial and cognitive considerations into organizational deployment is essential for strengthening sustainable utilization, especially in AI-centered initiatives. Moreover, refs. [22,23] emphasized the importance of understanding how price–value perceptions and managerial attitudes influence sustainable AI adoption throughout implementation journeys in business settings.

2. Literature Review

2.1. Review of Price–Value, Attitude Towards AI, Intention to Use and Sustainable Use of AI

Implementing artificial intelligence (AI) in organizational settings demands comprehension of fundamental constructs that affect its enduring deployment. This section examines the central concepts and their associations according to current scholarship.

Cost–benefit assessment represents the evaluative mechanism whereby organizations analyze anticipated AI technology benefits in relation to required financial investments for acquisition [24]. As institutional executives perceive that artificial intelligence returns surpass related costs, they exhibit heightened inclination toward deployment advocacy. Existing quantitative research demonstrates that value assessments significantly influence implementation readiness across various sectors, with [25] establishing favorable relationships between perceived worth and students’ generative AI utilization intentions in educational contexts. Likewise, ref. [26] discovered that perceived worth mediates the association between eHealth literacy and perspective toward AI in healthcare settings. Core results suggest that economic value assessments and institutional support demonstrate substantially favorable impact on enduring AI implementation among organizational executives, especially in technology-oriented enterprises [27]. This influence is reinforced by the results of [6], who illustrated that financial considerations and favorable perspectives are positively linked with sustained utilization and deployment effectiveness among technology-centered managers. However, ref. [28] identified a price–value paradox where practical barriers can reduce actual usage despite high perceived value, highlighting the complexity of this relationship.

AI-focused attitudes encompass managerial personnel’s comprehensive assessment orientations regarding artificial intelligence system implementation [29,30]. Such orientations develop from utility perceptions, usability considerations, and contextual organizational elements [31]. Contemporary research indicates that positive AI orientations become crucial for effective implementation results, with [32] determining associations between favorable AI viewpoints and institutional change preparedness in banking organizations. Ref. [33] determined that perspective toward AI is affected by social influence and perceived knowledge among educators, while [26] demonstrated that technology confidence and perceived worth serve mediating functions in forming attitudes toward AI in healthcare. Comprehensive research has revealed that financial considerations and perspectives substantially affect organizational managers’ deployment strategies, creating a favorable association with diminishing problems such as implementation resistance and technology apprehension [34]. Conversely, ref. [35] identified that perceived risk negatively impacts attitudes towards AI in e-commerce contexts.

Adoption willingness represents the extent to which an executive has developed deliberate plans to execute or avoid specific future actions concerning AI implementation [36]. This concept connects perspectives and actual utilization conduct. Current evidence suggests that behavioral intention to adopt AI is shaped by various elements including perceived utility, operational ease, and social standards. Ref. [37] discovered that perceived utility and social standards constitute important factors of AI implementation intention in accounting fields, while ref. [33] recognized performance expectancy and perspective toward AI as important predictors of behavioral intention in academic institutions. Ref. [38] illustrated that perceived advantages and awareness substantially influence AI implementation intention for energy management among academic faculty. Regarding organizational impact, ref. [39] documented that deployment success is favorably associated with executive perspectives and sustained utilization performance, with economic value mediating the relationship between these elements. Additionally, ref. [40] emphasized that user satisfaction serves as a crucial element in promoting AI adoption, mediating the positive effects of behavioral intention.

Enduring AI utilization denotes the sustained, responsible, and efficient incorporation of artificial intelligence systems into organizational operations while accounting for economic, social, and environmental consequences [41]. Current scholarship emphasizes that enabling conditions such as solid infrastructure and applied training are fundamental for sustainable AI implementation. Ref. [42] discovered that enabling conditions substantially correlate with enduring AI use among university faculty in Peru, while ref. [43] revealed that enabling conditions and effort expectancy significantly affect behavioral intention among academic librarians. Moral perceptions serve a crucial function in sustainable AI utilization, ref. [42] indicating that AI use is substantially connected to faculty ethical perceptions. Additionally, ref. [44] emphasized AI’s function in sustainability through ecological advantages, though efficient communication and promotion are required to harmonize public comprehension with technological progress. This relationship supplements the results of [45] who documented that economic value considerations are linked with enhanced sustainable utilization among technology-oriented managers, with perspectives mediating this influence.

Nevertheless, examining these constructs in the context of sustainable AI implementation faces substantial hurdles. Ref. [46] identified theoretical and methodological constraints in the existing AI adoption literature, which hinder the development of effective strategies for organizational deployment. In addition, sustainable use is often framed as an individual-level phenomenon, with insufficient attention to the organizational and economic conditions that shape its emergence among business managers [46].

Evidence from interventions in corporate settings is likewise encouraging. Ref. [47] showed that integrating a price–value analysis component into managerial AI training produces significant increases in sustainable use when implementation barriers arise. Even so, as ref. [48] emphasized, supporting business managers demands a comprehensive perspective, underscoring the need for multifaceted strategies to enhance adoption rates and sustainable use amid technological challenges.

2.2. Scientific Support of the Research Hypotheses

Quantitative research suggests that institutional executives displaying heightened financial value evaluations show improved implementation preparedness and technological satisfaction [49], reinforcing findings by [50], who established significant, favorable relationships between cost–benefit assessments and AI deployment willingness in technology-focused organizations. Studies have further determined that financial value evaluation functions as a core economic resource for guiding technology acceptance choices, contributing significantly to development of executive deployment strategies, principles, and technological goal orientation [51]. This influence is additionally reinforced by results from [6], who illustrated that the relationship between economic value and adoption willingness is mediated by perceived advantages and utility in institutional contexts. Furthermore, ref. [52] documented that all dimensions of economic value evaluation are inversely related to implementation resistance and deployment anxiety, accounting for substantial variance in organizational managers’ reluctance, uncertainty, and stress regarding AI adoption. The significance of economic value in AI deployment has been highlighted by [53], who recommended that business organizations should examine their function in fostering value-driven implementation strategies. These results are consistent with [54,55] who reported that the effect of price–value varies by organizational size and is linked to technological values, underscoring the complex, multidimensional character of this influence in business contexts. Consequently, the following is proposed:

Hypothesis 1.

The Price–Value significantly and positively influences the Intention to Use of artificial intelligence in business managers.

Research has demonstrated that executive perspectives are closely connected to technology implementation, deployment success, and institutional engagement [56,57] documented that favorable perspectives toward AI affect technology deployment decisions and operational efficiency. This influence is reinforced by results from [58], who recognized factors such as perceived utility, technological capability, organizational support, and executive confidence as important predictors of positive perspectives toward AI in business environments. The Technology Attitude Model (TAM) developed by [59] describes how perceived ease of use, perceived utility, social influence, and technological awareness contribute to forming positive perspectives and, subsequently, to organizational managers’ intention to adopt AI. Moreover, ref. [48] indicated that AI training and technology-centered interventions can diminish implementation resistance and deployment anxiety, thus enhancing usage intentions. The psychological mechanisms by which attitudes affect intention to use have been explained by [60] who proposed a model that includes technological self-efficacy, implementation goals, and adoption anxiety. These findings are supported by recent research from [61,62] which demonstrated that attitudes mediate the effects among managerial confidence, organizational culture, and implementation success and influence the relationship between AI training strategies and sustainable use. Therefore, we propose the following:

Hypothesis 2.

Attitude towards AI significantly and positively influences the Intention to Use of artificial intelligence in business managers.

This hypothesis is substantially reinforced by various investigations documenting the considerable influence of economic value on perspectives toward AI among organizational managers. Research has indicated that executives employ cost–benefit analysis procedures to assess AI deployment challenges, emphasizing economic value as a crucial determinant of technological perspectives [51]. This influence is additionally supported by findings from [9,63], who illustrated a positive relationship between economic value perceptions and perspectives toward AI in organizational contexts, suggesting that favorable economic value assessments indirectly diminish implementation resistance and strengthen positive perspectives. Research by [64,65] confirmed the positive association between price–value considerations and technological attitudes, whereas [66,67] explored how cost–benefit struggles can influence business managers’ technological adjustment and attitudes. Reference [68] reported that positive price–value perceptions are associated with lower levels of implementation stress, and ref. [69] reported that both economic considerations and value assessments significantly contribute to technology adoption attitudes. These results accord with [70,71], who documented associations between the level of price–value perception and attitudes toward emerging technologies. Furthermore, ref. [27] found that price–value considerations together with organizational support significantly foster favorable attitudes among senior managers, while ref. [72] underscored the pivotal role of price–value assessment in strengthening technological, cognitive, and operational attitudes, indicating its potential as a lever for shaping perspectives on AI implementation. Accordingly, the following is proposed:

Hypothesis 3.

The Price–Value significantly and positively influences the Attitude towards AI of artificial intelligence in business managers.

Investigations concerning adoption willingness and its impact on sustainable AI deployment discover considerable support in technology implementation scholarship. Research has demonstrated that adoption willingness favorably affects sustainable technology implementation among organizational managers [4], while stronger deployment intentions are linked with improved AI integration and institutional performance [49]. This influence is additionally reinforced by findings from [73], who established the relationship between usage intentions, technological beliefs, and sustainable AI practices among business organizations. In terms of organizational impact, research has shown that intention to use is positively correlated with sustainable implementation and long-term adoption among business professionals [74,75]. Ref. [76] emphasized the importance of clear implementation intentions and systematic adoption strategies among early-stage AI adopters, whereas ref. [77] reported that sustainable use levels vary by implementation phase, with organizations showing stronger intentions demonstrating greater sustainability in AI use than those with weaker intentions, suggesting a direct relationship between intention to use and sustainable AI implementation in business contexts. These findings are reinforced by recent studies in technology adoption. Ref. [78] found that intention to use explained 57% of the variance in sustainable AI implementation across diverse business sectors. Similarly, ref. [79] demonstrated through a longitudinal study that strong implementation intentions led to more sustainable and effective AI use patterns over time (β = 0.48, p < 0.001). Consequently, we propose the following hypothesis:

Hypothesis 4.

Intention to use has a significant and meaningful influence on the sustainable use of artificial intelligence in business managers.

Figure 1 presents the research model, which proposes the five hypotheses previously established.

Despite the growing body of literature on technology adoption, significant gaps remain in understanding how economic and psychological factors jointly influence sustainable AI implementation among business managers. While previous studies have examined price–value perceptions and attitudinal factors separately, no research has systematically integrated these constructs within a comprehensive framework that accounts for their interactive effects on behavioral intentions and sustained usage patterns. This gap is particularly salient in emerging economies, where resource constraints and nascent technological infrastructure create distinct adoption challenges that differ substantially from developed market contexts. Furthermore, the managerial perspective on AI adoption remains underexplored, despite managers’ pivotal role as decision-makers who shape organizational technology strategies and implementation outcomes.

The present study addresses these limitations by developing and empirically testing an integrated model that examines the sequential relationships among price–value perceptions, attitudes toward AI, intention to use, and sustainable AI implementation specifically among business managers in Peru. This research makes three distinct contributions. First, it advances theoretical understanding by demonstrating how economic evaluations influence adoption decisions through both direct and attitude-mediated pathways, thereby bridging rational choice and psychological perspectives on technology acceptance. Second, it extends the sustainability dimension of AI adoption research by establishing intention to use as a critical antecedent of long-term implementation patterns, moving beyond simplistic adoption-versus-rejection dichotomies. Third, it provides empirical evidence from an underrepresented Latin American context, illuminating how contextual factors shape technology adoption dynamics in resource-constrained business environments. The study’s practical implications include informing the design of AI implementation strategies that simultaneously address economic justification and psychological readiness, thereby enhancing adoption success rates and sustainability outcomes in organizational settings.

3. Materials and Methods

The hypotheses were tested through an empirical, quantitative study with a correlational design and explanatory orientation, applying systematic procedures to examine causal relationships among the variables. This approach was chosen to permit rigorous statistical evaluation of the proposed links between economic value, perspectives, adoption willingness, and sustainable AI use among organizational managers [80,81].

3.1. Details of Study Participants

The sample size of 390 organizational executives was established using non-probabilistic accessibility sampling, an approach chosen based on participant availability and the investigative nature of the research examining AI implementation relationships in organizational contexts. As indicated by [82], accessibility sampling is suitable for studies exploring associations between technological implementation variables in business contexts where availability and voluntary participation are fundamental. Sample size determination adhered to structural equation modeling requirements, as ref. [83] recommends 10–15 participants minimum per observed variable. Considering the 26 observed variables in the measurement model, the target sample surpassed the minimum requirement of 260 participants, finally achieving 390 participants to guarantee sufficient statistical power for PLS-SEM analysis.

It should be recognized that convenience sampling does not permit the calculation of traditional sampling error or the construction of confidence intervals, as would be possible with probabilistic designs. The non-probabilistic nature of the sample constitutes a salient limitation that restricts the generalizability of the results beyond the studied population of business managers in northern Peru. Although the sample size is adequate for the analytical techniques employed and provides sufficient statistical power for hypothesis testing [84], interpretation should remain bounded by this methodological constraint. The convenience-based approach limits the capacity to draw population-level inferences or to specify precise confidence levels, representing a key limitation to consider when assessing the study’s findings and their applicability to broader managerial populations.

As detailed in

Table 1. Sociodemographic characteristics of the sample (n = 390).

Characteristics	fi	%
Gender
Female	161	41.28
Male	229	58.72
Age group
25–29 years	60	15.38
30–39 years	176	45.13
40–49 years	110	28.21
50 or more years	44	11.28
Management Experience
Less than 5 years	80	20.51
5–10 years	152	38.97
11–15 years	100	25.64
More than 15 years	58	14.88

Note: fi = frequency.

, the gender distribution showed a predominance of male participants at 58.72% (229 participants) compared to 41.28% (161 participants) female. Regarding age distribution, the largest group was 30–39 years (45.13%, 176 participants), followed by 40–49 years (28.21%, 110 participants), and 25–29 years (15.38%, 60 participants). In terms of management experience, most participants had 5–10 years of experience (38.97%, 152 participants), followed by those with 11–15 years (25.64%, 100 participants), and those with less than 5 years (20.51%, 80 participants).

3.2. Instruments Used in the Study

Data gathering utilized a structured digital survey comprising four validated instruments, employing exclusively closed-ended questions with single-response alternatives to guarantee standardized measurement and enable statistical analysis. The survey structure contained three primary sections: informed consent (1 item), sociodemographic characteristics (3 items), and construct measurement scales (26 items).

Four established measurement instruments enabled information collection:

• Price–Value Scale (PV): The empirically confirmed instrument developed by ref. [24] was employed to evaluate managerial perspectives regarding AI value propositions in relation to financial requirements. This four-indicator instrument demonstrates solid statistical properties within technology acceptance research, including components addressing cost–benefit evaluation, value justification, investment returns, and institutional benefit delivery.
• Attitude towards AI Scale (ATTAI): The adapted version by ref. [80] was employed to evaluate executive personnel’s comprehensive evaluative viewpoints regarding AI implementation. This nine-indicator scale exhibits excellent reliability within institutional contexts, encompassing dimensions including perceived benefits, AI utilization ease, professional efficiency improvement, and trust in AI capabilities.
• Intention to Use Scale (INUS): The version confirmed by ref. [78] was utilized to assess executives’ behavioral intentions concerning AI implementation. This 6-item instrument has shown high internal consistency, examining planning to employ AI, deployment intentions, integration objectives, adoption readiness, and productivity improvement goals.
• Sustainable Use of AI Scale (SUSAI): The modification by ref. [85] was employed to evaluate enduring, responsible AI deployment. This 7-item scale has exhibited solid psychometric properties in organizational settings, assessing responsible utilization commitment, long-term impact evaluation, institutional benefit integration, balanced usage preservation, sustainable value creation, and consistency in beneficial implementation.

All instruments used a 5-point Likert scale ranging from (1) “strongly disagree” to (5) “strongly agree.” The questionnaire was distributed via a secure digital platform (SurveyMonkey), ensuring data privacy and accessibility across different organizational contexts.

Pilot Testing and Validation: Prior to main data collection, a pilot study was conducted with 25 business managers from 5 organizations not included in the final sample. The pilot testing revealed high internal consistency (Cronbach’s α > 0.80 for all scales) and identified minor wording adjustments to enhance clarity in the Peruvian business context. Participants in the pilot study suggested simplifying technical terminology in two items, which were subsequently modified while maintaining construct validity.

Sampling Strategy: Organizations were selected through purposive sampling based on three criteria: (1) active AI implementation or consideration, (2) willingness to participate in research, and (3) availability of middle to senior management personnel. Initial contact was established through professional networks and chamber of commerce databases in northern Peru. From 45 organizations contacted, 28 agreed to participate, representing a 62% organizational response rate. The sample size calculation was based on structural equation modeling requirements, with a minimum of 15 participants per observed variable, resulting in a target sample of 390 participants across participating organizations.

The complete questionnaire, including all items and their psychometric properties, is provided in

Table A1. Complete item inventory with factor loadings and validation results.

Construct	Item Code	Item Description	Factor Loading	Standard Deviation (STDEV)	p Values	Status	AVE
Price–Value (PV)	PV1	The value obtained from AI usage justifies its cost	0.734	0.023	0.000	Retained	0.562
	PV2	AI provides good value considering the price to be paid	0.756	0.045	0.000	Retained
	PV3	The benefits provided by AI compensate for the required investment	0.835	0.053	0.000	Retained
	PV4	At its current price point, AI delivers good value for my organization	0.872	0.024	0.000	Retained
Attitude Towards AI (ATTAI)	ATTAI1	I consider AI usage beneficial for my work	0.869	0.060	0.000	Retained	0.678
	ATTAI2	I maintain a positive attitude towards AI implementation in my professional activities	0.832	0.039	0.000	Retained
	ATTAI3	I believe implementing AI is a judicious decision	0.831	0.051	0.000	Retained
	ATTAI4 *	I feel that AI poses risks to data security in my organization	0.428	0.067	0.142	Eliminated
	ATTAI5	I feel comfortable with the prospect of utilizing AI in my work	0.778	0.033	0.000	Retained
	ATTAI6	I believe AI can significantly enhance my professional performance	0.989	0.048	0.000	Retained
	ATTAI7	I value the capabilities that AI can contribute to my organization	0.705	0.037	0.000	Retained
	ATTAI8	I am enthusiastic about the prospect of working with AI	0.880	0.051	0.000	Retained
	ATTAI9	I trust that AI can help me perform my tasks more effectively	0.859	0.027	0.000	Retained
Intention to Use (INUS)	INUS1	I plan to utilize AI in my work when it becomes available	0.754	0.056	0.000	Retained	0.654
	INUS2	I intend to implement AI in my professional activities	0.867	0.027	0.000	Retained
	INUS3 *	I would recommend AI adoption to colleagues in my organization	0.471	0.072	0.089	Eliminated
	INUS4	I aim to integrate AI into my work processes	0.769	0.037	0.000	Retained
	INUS5	I am willing to adopt AI in my professional practice	0.876	0.039	0.000	Retained
	INUS6	I intend to utilize AI to enhance my productivity	0.754	0.059	0.000	Retained
Sustainable Use of AI (SUSAI)	SUSAI1	I plan to use AI in a responsible and sustainable manner	0.873	0.043	0.000	Retained	0.712
	SUSAI2	I am committed to implementing AI while considering its long-term impact	0.911	0.032	0.000	Retained
	SUSAI3	I intend to integrate AI in a way that benefits the entire organization	0.823	0.062	0.000	Retained
	SUSAI4 *	I will monitor the environmental impact of AI technologies in my work	0.463	0.081	0.116	Eliminated
	SUSAI5	I will seek to maintain a balanced and effective use of AI	0.842	0.032	0.000	Retained
	SUSAI6	I will strive to utilize AI in a way that generates sustainable value	0.864	0.046	0.000	Retained
	SUSAI7	I will ensure that AI usage remains consistent and beneficial over time	0.876	0.053	0.000	Retained

* Note: All items use a 5-point Likert scale where 1 = Strongly disagree; 2 = Disagree; 3 = Neither agree nor disagree; 4 = Agree; 5 = Strongly agree. Items marked with an asterisk () were eliminated during confirmatory factor analysis. Elimination criterion: factor loading < 0.50 or p-value > 0.05, following recommendations by Hair et al. (2019) [83].

and

Table A2. Summary of construct validation and item retention.

Construct	Initial Items	Eliminated Items	Final Items	Items Retained (%)	Cronbach’s α	CR (rho_c)	AVE
Price–Value (PV)	4	0	4	100%	0.856	0.896	0.562
Attitude Towards AI (ATTAI)	9	1 (ATTAI4)	8	88.9%	0.822	0.892	0.678
Intention to Use (INUS)	6	1 (INUS3)	5	83.3%	0.742	0.752	0.654
Sustainable Use of AI (SUSAI)	7	1 (SUSAI4)	6	85.7%	0.811	0.858	0.712
Total	26	3	23	88.5%	-	-	-

Note: All reliability coefficients exceed the minimum acceptable threshold of 0.70 [95]. Average variance extracted (AVE) values surpass the 0.50 criterion, confirming adequate convergent validity [86].

(Appendix A) for transparency and replication purposes. Table A1 presents the complete item inventory with factor loadings and validation results, while Table A2 summarizes construct validation and item retention rates.

3.3. Procedure and Data Analysis

Data collection took place from June to September 2024 in multiple business organizations across northern Peru. After securing the requisite institutional authorizations, the instrument was disseminated via corporate email and professional networks to organizational managers, accompanied by a detailed description of the study’s objectives regarding AI implementation and the voluntary nature of participation.

To promote consistency across organizational contexts, a unified data-collection protocol was implemented, incorporating standardized participant instructions and explicit guidelines to uphold data quality in virtual business settings.

The analytical sequence comprised: (1) data cleaning and preparation in Microsoft Excel; (2) descriptive statistics to characterize the sample; and (3) confirmatory factor analysis (CFA) to assess convergent validity through factor loadings and average variance extracted (AVE), with thresholds of 0.70 and 0.50, respectively. During CFA, items ATTAI4, INUS3, and SUSAI4 were removed due to insufficient loadings (<0.50). Reliability was evaluated using Cronbach’s alpha and composite reliability (CR), while discriminant validity was examined with the [86] criterion and the heterotrait–monotrait (HTMT) ratio.

Finally, hypothesis testing was conducted via partial least squares structural equation modeling (PLS-SEM) in SMART-PLS v.4.0, estimating relationships among price–value, attitudes, intention to use, and sustainable AI implementation in business contexts.

4. Results

4.1. Results of the Measurement Model

Partial least squares structural equation modeling (PLS-SEM) was applied, and convergence of the measurement model was evaluated through confirmatory factor analysis (CFA). As indicated in

Table 2. Results of the confirmatory factor analysis-CFA.

Items		Factor Loading	Standard Deviation (STDEV)	p Values	AVE	Construct
The value obtained from AI usage justifies its cost	PV1	0.734	0.023	0.000	0.562	Price–Value (PV)
AI provides good value considering the price to be paid	PV2	0.756	0.045	0.000
The benefits provided by AI compensate for the required investment	PV3	0.835	0.053	0.000
At its current price point, AI delivers good value for my organization	PV4	0.872	0.024	0.000
I consider AI usage beneficial for my work	ATTAI1	0.869	0.060	0.000	0.678	Attitude towards AI (ATTAI)
I maintain a positive attitude towards AI implementation in my professional activities	ATTAI2	0.832	0.039	0.000
I believe implementing AI is a judicious decision	ATTAI 3	0.831	0.051	0.000
I feel comfortable with the prospect of utilizing AI in my work	ATTAI 5	0.778	0.033	0.000
I believe AI can significantly enhance my professional performance	ATTAI 6	0.989	0.048	0.000
I value the capabilities that AI can contribute to my organization	ATTAI 7	0.705	0.037	0.000
I am enthusiastic about the prospect of working with AI	ATTAI 8	0.880	0.051	0.000
I trust that AI can help me perform my tasks more effectively	ATTAI 9	0.859	0.027	0.000
I plan to utilize AI in my work when it becomes available	INUS1	0.754	0.056	0.000	0.654	Intention to Use (INUS)
I intend to implement AI in my professional activities	INUS2	0.867	0.027	0.000
I aim to integrate AI into my work processes	INUS4	0.769	0.037	0.000
I am willing to adopt AI in my professional practice	INUS5	0.876	0.039	0.000
I intend to utilize AI to enhance my productivity	INUS6	0.754	0.059	0.000
I plan to use AI in a responsible and sustainable manner	SUSAI1	0.873	0.043	0.000	0.712	Sustainable use of AI (SUSAI)
I am committed to implementing AI while considering its long-term impact	SUSAI2	0.911	0.032	0.000
I intend to integrate AI in a way that benefits the entire organization	SUSAI3	0.823	0.062	0.000
I will seek to maintain a balanced and effective use of AI	SUSAI5	0.842	0.032	0.000
I will strive to utilize AI in a way that generates sustainable value	SUSAI6	0.864	0.046	0.000
I will ensure that AI usage remains consistent and beneficial over time	SUSAI7	0.876	0.053	0.000

Note: All items use a 5-point Likert scale where 1 = Strongly disagree; 2 = Disagree; 3 = Neither agree nor disagree; 4 = Agree; 5 = Strongly agree.

, all item loadings are ≥0.70, meeting adequacy standards per [87]. Moreover, the average variance extracted (AVE) for every construct exceeds the 0.50 threshold, in line with [88].

Table 3. Reliability, discriminant validity, and coefficients of determination.

Construct	α	CR(rho_a)	CR(rho_c)	R2	Q2 Predict	HAP	RES	SPI	HTMT
INUS	0.742	0.732	0.752	0.892	0.899	0.776			0.402
SUSAI	0.811	0.831	0.858	0.851	0.979	0.430	0.874		0.251
ATTAI	0.822	0.839	0.892	-	-	0.598	0.374	0.781	0.288
PV	0.856	0.873	0.896

Note: Diagonal values (in bold) represent the square root of AVE for each construct. Off-diagonal values represent inter-construct correlations. INUS = Intention to Use; SUSAI = Sustainable Use of AI; ATTAI = Attitude Towards AI; PV = Price–Value; α = Cronbach’s alpha; CR = Composite Reliability; R² = Coefficient of determination; Q² = Predictive relevance; HTMT values (not shown) were below 0.85 threshold, confirming discriminant validity.

reports the constructs’ reliability and discriminant validity. Reliability was assessed via Cronbach’s alpha (α) and composite reliability (CR), including rho_a and rho_c. In line with [80,83] coefficients above 0.70 denote acceptable reliability; as indicated in Table 3, all constructs surpass this benchmark, evidencing adequate measurement consistency.

Regarding explanatory power (R²), PV and ATTAI explain 89.2% of the variance in INUS, while INUS accounts for 85.1% of the variance in SUSAI, demonstrating substantial explanatory capacity within the framework. Discriminant validity was examined using the [86] criterion, which requires that the square root of the average variance extracted (AVE)—the diagonal entries—exceed the inter-construct correlations appearing off the diagonal in the corresponding rows and columns. As shown in Table 3, all constructs meet this condition, confirming discriminant validity. Additionally, heterotrait–monotrait (HTMT) ratios are below the 0.85 threshold recommended by [89], further reinforcing the instrument’s discriminant validity.

Model-fit indices are pivotal for establishing the convergent validity of the measurement model [90] and for assessing the alignment between observed data and theoretical expectations [87]. As summarized in

Table 4. Model fit.

Criteria	Estimated Model	Threshold	Author	Decision
SRMR	0.065	<0.85	[91]	Acceptable
d_ULS	2.223
d_G	0.672
χ2/df	1.342	Between 1 and 3	[92]	Acceptable
NFI	1.267	>0.90	[92]	Acceptable

, the standardized root mean square residual (SRMR) is 0.065, falling within the acceptable bounds recommended by [91]. The chi-square to degrees-of-freedom ratio (χ²/df) is 1.342, within the suggested 1–3 interval and indicative of satisfactory fit according to [92]. Additionally, the normalized fit index (NFI) is 1.267, exceeding the minimum 0.90 threshold noted by [92]. Taken together, these indicators suggest that the measurement model is adequate and consistent with the data.

4.2. Testing the Research Hypotheses

Table 5. Results of hypothesis testing.

Hypothesis		β	f2	p Value	Percentile		SD	Decision
					2.50%	97.50%
H1	PV → INUS	0.662 ***	0.670	0.000	0.444	0.799	0.065	Accepted
H2	ATTAI →INUS	0.456 ***	0.562	0.000	0.360	0.657	0.051	Accepted
H3	PV→ ATTAI	0.374 ***	0.452	0.000	0.275	0.507	0.060	Accepted
H4	INUS → SUSAI	0.749 ***	0.765	0.000	0.552	0.892	0.062	Accepted

Note. β = path coefficient; SE = standard deviation; *** p < 0.001; interpretation of effect size according to [84]: small f² < 0.02; medium f² < 0.15; large f² ≥ 0.35.

and Figure 2 summarize the hypothesis-testing outcomes within the proposed framework. As shown in Table 5, all hypotheses received empirical support with path coefficients ranging from β = 0.374 to β = 0.749, standard deviations between 0.051 and 0.065, and 95% confidence intervals that exclude zero across all relationships. First, H1 (see Figure 2) indicates a substantial effect of Price–Value (PV) on Intention to Use (INUS), with a path coefficient of β = 0.662 (p < 0.001, SD = 0.065, 95% CI [0.444, 0.799]). This suggests that perceptions of value relative to price directly and favorably influence organizational managers’ intention to adopt artificial intelligence (AI). The effect size (f² = 0.670) was assessed following [84] guidelines and categorized as large.

For H2 (see Figure 2), Attitude toward AI (ATTAI) exerts a substantial effect on INUS (β = 0.456, p < 0.001, SD = 0.051, 95% CI [0.360, 0.657]), with an effect size of f² = 0.562, categorized as large. This finding indicates that attitudinal evaluations play a decisive role in adoption intention, underscoring the importance of psychological factors in technology implementation.

In turn, H3 (see Figure 2) shows that PV positively and substantially affects ATTAI, with a path coefficient of β = 0.374 (p < 0.001, SD = 0.060, 95% CI [0.275, 0.507]) and an effect size of f² = 0.452, interpreted as moderate-to-large. This implies that perceived economic value directly shapes favorable attitudes toward AI, potentially reducing psychological barriers to implementation. Finally, H4 is supported (see Figure 2), revealing a substantial effect of INUS on Sustainable Use of AI (SUSAI) (β = 0.749, p < 0.001, SD = 0.062, 95% CI [0.552, 0.892]) with a large effect size (f² = 0.765). This highlights that intention to use strongly and directly influences sustainable use, emphasizing the importance of cultivating intention as a critical step toward sustainable AI practices.

5. Discussion

The results provide substantial empirical support for the relationships among cost–benefit evaluations, attitudes, implementation intentions, and sustained AI use among business executives. Testing of the proposed hypotheses yielded notable outcomes that deepen understanding of these constructs within corporate technology-implementation contexts.

For the first hypothesis, the evidence confirms that economic value has a significant positive effect on willingness to adopt artificial intelligence among organizational managers (β = 0.662, p < 0.001). This strong association indicates that when AI deployment is perceived to offer a favorable cost–benefit balance, managers are more likely to develop robust intentions to implement the technology. This aligns with [49], which reported that positive economic-value perceptions predict higher implementation intentions (β = 0.45, p < 0.001). Similarly, ref. [50] found that price–value considerations significantly influence implementation intentions across varied organizational settings (β = 0.52, p < 0.01). The magnitude of this relationship is consistent with [78], identifying price–value as a core economic driver of managerial adoption decisions. Moreover, ref. [6] demonstrated that the impact of price–value on intention to use is mediated by perceived benefits in organizational environments (β = 0.48, p < 0.001), further corroborating these findings. These results are reinforced by [79], who established that organizational readiness—including economic considerations—significantly influences the AI adoption journey, with technology readiness levels serving as critical benchmarks for successful implementation across different organizational phases.

For the second hypothesis, the findings illustrate that perspectives toward AI substantially affect adoption willingness (β = 0.456, p < 0.001). This influence indicates that positive executive perspectives toward AI technology serve a vital function in forming deployment intentions. This result is aligned with [56] research demonstrating that executive perspectives strongly predict technology implementation intentions (β = 0.43, p < 0.001). Reference [57] similarly found that positive attitudes towards AI significantly influence implementation decisions (β = 0.39, p < 0.01). Our results also align with [58] identification of attitudinal factors as significant predictors of AI adoption intentions. Furthermore, ref. [48] demonstrated that positive attitudes reduce adoption resistance and enhance implementation intentions (β = 0.51, p < 0.001), supporting the importance of attitudinal factors in technology adoption. This finding gains additional support from [80], who documented that computer skills training—which shapes technological attitudes—demonstrates significant positive effects on organizational performance and digital transformation readiness in Eastern European companies, suggesting that attitude formation through skill development constitutes a critical pathway for sustainable technology adoption.

The third hypothesis—stating that price–value significantly influences attitudes toward AI—was supported (β = 0.374, p < 0.001). This association indicates that favorable price–value perceptions foster more positive attitudes toward AI implementation. Consistent evidence appears in [9], which found that price–value considerations significantly predict technological attitudes (β = 0.42, p < 0.001). Similarly, ref. [64] reported comparable effects for cost–benefit perceptions and technology attitudes (β = 0.38, p < 0.01). In addition, ref. [66] showed that price–value assessments shape technological attitudes in business contexts (β = 0.45, p < 0.001), and ref. [68] further demonstrated that positive price–value perceptions reduce implementation stress and enhance attitudes toward new technologies (β = 0.47, p < 0.001). The relationship between economic evaluation and attitude formation is further illuminated by [79], who argued that successful AI adoption requires alignment between technology readiness and organizational preparedness across people, processes, and data dimensions, with economic considerations serving as foundational elements in this alignment process.

Finally, the fourth hypothesis—assessing the effect of intention to use on sustainable AI implementation—was strongly supported (β = 0.749, p < 0.001). This robust association indicates that managers’ implementation intentions are pivotal determinants of sustainable AI use. Convergent evidence appears in [4], which reported that implementation intentions strongly predict sustainable technology adoption (β = 0.56, p < 0.001); likewise, ref. [73] documented a significant link between usage intentions and sustainable practices (β = 0.61, p < 0.001). In the same direction, ref. [74] found a positive correlation between intention to use and long-term technology adoption (β = 0.53, p < 0.001), and ref. [78] showed that implementation intentions account for a substantial share of variance in sustainable AI use across business sectors (β = 0.58, p < 0.001). This relationship is contextualized by [93], who emphasized that sustainable AI implementation requires progression through multiple technology readiness levels, with intention serving as the bridge between initial adoption decisions and long-term operational success. Additionally, ref. [94] demonstrated that organizational investments in digital competencies—reflecting strong implementation intentions—correlate with improved performance outcomes across multiple economic indicators, reinforcing the critical role of intentionality in achieving sustainable technology integration.

The interpretation of these findings must account for contextual factors specific to northern Peru that may influence the observed relationships. The region’s digital infrastructure—characterized by variable internet connectivity, limited cloud computing resources, and nascent AI service ecosystems—likely shapes how business managers evaluate price–value propositions and form implementation intentions. Organizations operating in contexts with underdeveloped technological infrastructure face higher implementation costs and greater uncertainty regarding return on investment, potentially amplifying the importance of clear economic value demonstrations. Additionally, Peru’s economic environment, marked by small and medium enterprise prevalence, resource constraints, and cautious investment cultures, may heighten sensitivity to cost–benefit considerations relative to contexts where capital availability permits more experimental technology adoption. The regional business culture, which emphasizes relationship-based decision-making, hierarchical organizational structures, and collective rather than individual risk assessment, could moderate how attitudes translate into intentions and subsequently into sustained usage. These contextual elements suggest that the particularly strong relationship between price–value and intention observed here (β = 0.662) may reflect situational factors where economic justification assumes paramount importance due to resource scarcity and risk aversion. In more developed markets with mature AI ecosystems, robust digital infrastructure, and risk-tolerant business cultures, the relative influence of economic versus attitudinal factors might differ, with psychological elements potentially assuming greater prominence. Future research examining these relationships across diverse economic and cultural contexts would illuminate the boundary conditions for the patterns identified in this Peruvian sample and advance toward culturally contextualized models of AI adoption.

The credibility of the findings is bolstered by strong psychometric evidence, with model fit indices meeting established thresholds (SRMR = 0.065; NFI = 1.267). Additionally, high reliability and discriminant validity—indicated by HTMT values below 0.85—support the conclusion that the observed relationships represent genuine patterns within the studied population of business managers.

5.1. Theoretical Implications

This research advances the theoretical understanding of AI adoption in organizational contexts by establishing an integrated framework that bridges economic and psychological perspectives. The model’s substantial explanatory power (R² = 0.892 for intention to use; R² = 0.851 for sustainable use) demonstrates that combining price–value perceptions with attitudinal constructs provides a more comprehensive account of technology adoption than models examining these factors in isolation.

The study makes three significant theoretical contributions. First, it validates the mediating role of attitudes in the relationship between economic evaluations and behavioral intentions, confirming that price–value perceptions influence adoption decisions both directly and indirectly through attitude formation. This dual-pathway mechanism enriches technology acceptance frameworks by demonstrating how rational economic assessments interact with psychological factors to shape implementation decisions. Second, the research extends sustainable technology adoption theory by establishing intention to use as a critical antecedent of sustainable AI implementation (β = 0.749, p < 0.001), providing empirical evidence that volitional commitment precedes enduring usage patterns. Third, the framework contributes to business management literature by demonstrating that AI adoption among managers follows distinct patterns characterized by strong economic rationality (β = 0.662 for price–value on intention) combined with attitudinal mediation, suggesting that managerial technology adoption differs from end-user acceptance in emphasizing cost–benefit considerations.

The integration of these constructs addresses a theoretical gap identified in prior research, where economic and psychological factors were examined separately despite their evident interdependence in organizational decision-making. The robust path coefficients and high explanatory variance indicate that this integrated approach captures essential mechanisms underlying sustainable AI adoption in business environments.

5.2. Practical Implications

The findings yield actionable insights for organizational leaders and policymakers seeking to promote sustainable AI implementation. The pronounced effect of price–value on both attitudes (β = 0.374) and intentions (β = 0.662) underscores the necessity of clearly articulating AI’s economic value proposition to management teams. Organizations should develop comprehensive cost–benefit analyses that quantify return on investment, productivity gains, and competitive advantages, presenting these metrics through structured business cases that resonate with managerial decision-making frameworks.

The substantial influence of attitudes on implementation intentions (β = 0.456) suggests that cultivating favorable managerial perspectives toward AI constitutes a strategic priority. Organizations can achieve this through targeted interventions including hands-on demonstration projects, pilot programs that showcase AI capabilities in relevant business contexts, and structured change management initiatives that address concerns while building confidence. Training programs should emphasize not only technical competencies but also strategic understanding of AI’s organizational implications, enabling managers to develop informed, positive attitudes grounded in practical knowledge.

The strong association between intention and sustainable use (β = 0.749) highlights that genuine commitment—rather than superficial compliance—drives lasting implementation success. Organizations should foster authentic engagement through participatory implementation processes, where managers contribute to AI strategy formulation and deployment planning. This approach transforms passive recipients into active stakeholders, strengthening implementation intentions and increasing the likelihood of sustained adoption.

For policymakers and industry associations, these findings suggest that AI adoption initiatives should incorporate economic literacy components alongside technical training, recognizing that managers require clear understanding of both financial implications and technological capabilities. Support mechanisms might include frameworks for assessing AI readiness, industry-specific cost–benefit benchmarks, and knowledge-sharing platforms that facilitate peer learning among business leaders navigating AI implementation.

6. Limitations and Future Research

This study presents several constraints that influence result interpretation and establish directions for future inquiry. The primary limitation concerns geographic scope, as the sample of 390 organizational managers from northern Peru restricts generalizability to other cultural and economic contexts. Northern Peru’s particular business environment—characterized by specific technological infrastructure, regulatory frameworks, and organizational cultures—may shape the relationships observed among price–value perceptions, attitudes, and AI adoption in ways that differ from other regions. The applicability of these findings to managers in developed economies, other Latin American countries, or Asian and African markets remains uncertain and warrants cautious interpretation.

The cross-sectional design constitutes a second significant limitation, precluding definitive causal inferences regarding the relationships among constructs. While structural equation modeling enables assessment of directional relationships, the temporal ordering of variables cannot be conclusively established without longitudinal data. Future research should employ panel designs tracking managers across multiple time points to examine how price–value perceptions, attitudes, and intentions evolve during AI implementation journeys, and to establish temporal precedence among these constructs.

Methodologically, exclusive reliance on self-report measures introduces potential common method bias, despite statistical controls. Managers may overestimate their implementation intentions or provide socially desirable responses regarding attitudes toward emerging technologies. Future studies should incorporate multi-method approaches combining self-reports with objective behavioral indicators such as actual AI adoption rates, implementation timelines, and resource allocation decisions. Additionally, collecting data from multiple organizational stakeholders—including IT personnel, frontline employees, and senior executives—would provide triangulated perspectives on AI implementation processes.

The non-probabilistic sampling approach, while appropriate for this exploratory investigation, limits statistical generalization to broader managerial populations. Future research should employ probability sampling methods across diverse organizational types, industry sectors, and geographic regions to enhance external validity. Comparative studies examining AI adoption patterns across cultural contexts would illuminate whether the relationships identified here represent universal patterns or context-specific phenomena.

Several theoretical constructs potentially relevant to sustainable AI adoption were not examined in this study. Organizational culture, leadership support, technological infrastructure readiness, and regulatory environments likely moderate or mediate the relationships identified here. Future investigations should develop more comprehensive models incorporating these organizational and contextual factors. Additionally, examining potential curvilinear relationships—such as whether extremely high price–value perceptions produce diminishing returns on attitudes or intentions—would refine theoretical understanding.

The brief Likert scales employed, while psychometrically sound, may not capture the full complexity of constructs such as ethical considerations in AI use, long-term sustainability commitment, or organizational change readiness. Future research might employ mixed methods designs combining quantitative surveys with qualitative interviews to explore nuanced aspects of managerial AI adoption that standardized scales cannot adequately assess. Longitudinal case studies tracking organizations through complete AI implementation cycles would provide rich insights into temporal dynamics and implementation challenges.

Finally, the model does not address implementation outcomes beyond sustainable use intentions. Future research should examine relationships between sustainable AI use and organizational performance indicators such as productivity gains, competitive advantage, innovation capacity, and financial returns. Understanding these downstream consequences would complete the causal chain from initial price–value perceptions through sustainable implementation to organizational outcomes, providing a more complete account of AI adoption’s organizational impact.

7. Conclusions

This investigation establishes that sustainable AI implementation among business managers emerges from the coordinated influence of economic value perceptions, favorable attitudes, and strong implementation intentions. The research demonstrates that these factors operate as an integrated system rather than isolated variables, with economic evaluations shaping psychological responses that subsequently influence behavioral commitments leading to sustained usage patterns.

Three principal findings merit emphasis. First, price–value perceptions exert both direct influence on implementation intentions and indirect effects mediated by attitudinal formation, challenging simplistic models that treat cost–benefit analysis and psychological acceptance as separate processes. Second, the robust relationship between implementation intentions and sustained use patterns confirms that volitional engagement, rather than mandated compliance, drives lasting technology integration. Third, the model’s substantial explanatory power validates the theoretical premise that managerial technology adoption reflects complex interplay between rational calculation and affective response.

These findings yield actionable implications for organizational practice. Successful AI implementation strategies must simultaneously address economic justification through comprehensive cost–benefit analyses and psychological readiness through experiential learning and participatory decision-making. Organizations that invest substantial effort in building authentic intentions during pre-implementation phases establish stronger foundations for sustained adoption than those emphasizing post-implementation support alone.

The study acknowledges contextual specificity as both a limitation and opportunity for future research. While the patterns observed among Peruvian business managers align with the literature on broader technology acceptance, cultural, economic, and institutional factors specific to emerging economies may influence the relative weight of economic versus psychological determinants. Comparative research examining these relationships across diverse contexts would illuminate boundary conditions and advance toward more nuanced, context-sensitive models of organizational technology adoption. Future investigations should also employ longitudinal designs to establish temporal precedence among constructs and incorporate objective behavioral indicators alongside self-report measures to strengthen causal inferences.