Potential domains, challenges and evaluation standards for integrating artificial intelligence and BIM into construction processes Open Access

Recent studies (Bassir et al., 2023; Heidari et al., 2024) have highlighted growing interest in the integration of artificial intelligence (AI) techniques with building information modeling (BIM). Building on these developments, researchers have explored the combined use of BIM and AI (BIM–AI) for tasks such as automated site progress monitoring, clash detection and construction simulation (Saleh et al., 2024; Dolhopolov et al., 2024; Andritsou et al., 2024). This integration is seen as a way to enhance the capabilities of each system while addressing the limitations of BIM alone (Adetayo and Innocent, 2022; Albert et al., 2022; Rangasamy and Yang, 2024). According to Albert et al. (2022), AI’s ability to analyze BIM data supports more informed decision-making and optimizes project delivery. Despite its potential, the integration of BIM and AI presents significant operational challenges that may hinder its adoption and effectiveness. Understanding these challenges is critical to anticipating and mitigating disruptions, thereby preserving the resilience and performance of BIM–AI systems. Without addressing operational concerns, construction projects risk inefficiencies, increased risks and failure in solving complex problems.

In addition to identifying operational barriers, establishing a framework for evaluating integrated BIM–AI systems is essential. As emphasized by Varun et al. (2018) and Urbieta et al. (2023), having standardized assessment criteria helps set benchmarks for performance evaluation. These criteria enable stakeholders to assess the effectiveness of AI models within BIM environments, refine performance and enhance the overall quality of outcomes. Evaluation standards also provide clear performance objectives and facilitate continuous improvement, thereby promoting consistency, adaptability and client satisfaction in an evolving construction landscape. Consequently, identifying operational issues and developing evaluation criteria are critical steps for the effective integration of AI in BIM-based construction operations. While prior research has examined the benefits and applications of BIM–AI, there remains a lack of conceptual clarity regarding specific domains of integration within construction workflows. Moreover, comprehensive investigations into the operational challenges and the absence of standardized assessment frameworks underscore a significant research gap.

This study addresses these gaps by exploring the potential domains, operational barriers and evaluation standards necessary for the effective integration of AI and BIM in construction processes. To address these issues, this study is guided by the following research questions: What are the potential domains for integrating AI into construction operations using BIM? What are the key operational challenges faced when implementing BIM–AI in construction organizations? What evaluation standards can be used to measure the performance and effectiveness of integrated BIM–AI systems?

Historically, construction organizations have relied on BIM tools and platforms to support a wide range of operational tasks, including project management, documentation, quantity take-off, cost estimation, schedule tracking and risk mitigation (Shaqour, 2022). While these tools offer substantial benefits in streamlining construction processes, they are not without limitations. Notably, a lack of automation and the persistence of non-value-adding activities within BIM workflows can hinder efficiency. If left unaddressed, these drawbacks may diminish the net value of BIM in construction operations (Samad et al., 2022). To overcome these limitations, researchers and industry professionals have increasingly turned to AI as a complementary technology. AI offers the potential to automate repetitive tasks, optimize resource usage, enhance decision-making and eliminate inefficiencies. AI encompasses a broad range of subfields, including machine learning, fuzzy logic, neural networks, genetic algorithms, knowledge-based systems, computer vision, robotics and optimization. These techniques can significantly improve operational performance through process automation, enhanced risk management, safety monitoring and predictive analytics (Sofiat et al., 2022; Justus, 2023; Ghasemi Poor Sabet and Heap-Yih, 2020).

The integration of AI with BIM – commonly referred to as BIM–AI – has enabled a variety of innovative applications in construction organizations. These include automated progress tracking, clash detection, predictive cost control, intelligent scheduling, design optimization and real-time safety monitoring. Studies such as those by Tan et al. (2022) have emphasized how such applications improve productivity, minimize rework, reduce delays and cost overruns and enhance project management outcomes. Critically, the literature reveals a growing awareness of the potential domains for BIM–AI integration. Albert et al. (2022) emphasize the importance of interdisciplinary synergy, advocating for integration with emerging technologies like the internet of Things (IoT), blockchain and Geographic Information Systems (GIS) to enrich BIM models with dynamic, real-time data. Others, such as Pan and Zhang (2023), Rane (2023) and Adeel et al. (2023), demonstrate how AI-enhanced BIM can address key industry challenges – such as rework, structural damage identification and energy efficiency.

Further, Heidari et al. (2024) argue for the comprehensive digitalization of the construction lifecycle by integrating AI and other intelligent systems into BIM environments. This is echoed by studies such as Saleh et al. (2024), who explore sustainability-focused applications, including renewable energy integration and intelligent building controls. Dolhopolov et al. (2024) introduce the concept of an “evolutionary digital duplicate” using AI and BIM, pushing the boundaries toward digital twin development. Alavi et al. (2024) extend this by combining AI – specifically particle swarm optimization – with BIM and digital twin systems for facility layout optimization, underscoring the synergistic potential of these technologies in design and operations. Zhao et al. (2025) apply BIM–AI integration to disaster resilience, specifically assessing seismic vulnerability in urban infrastructure. Their findings confirm the value of integrated systems in risk prediction and pre-crisis planning, thereby expanding BIM–AI’s scope to urban-scale planning and infrastructure resilience.

These studies contribute to BIM–AI integration in construction operations by underscoring the current inefficiencies in BIM use, such as lack of automation and manual workflows; articulating how AI enhances BIM through automation, optimization and intelligent decision-making; showing how integrating AI with other technologies (IoT, blockchain, digital twins) within BIM platforms opens up new operational domains and pointing to the lack of empirical, practice-based data and standardized evaluation frameworks, setting the stage for further applied research. From a critical perspective, these studies collectively demonstrate that the integration of AI into BIM is not merely an enhancement of existing tools but a fundamental shift toward intelligent, data-driven construction ecosystems. The adoption of AI-driven methodologies transforms BIM from a static modeling tool into a dynamic decision-support system. However, while academic insights provide promising theoretical frameworks, practical implementation remains fragmented due to challenges such as data standardization, interoperability, workforce readiness and a lack of unified evaluation frameworks. Therefore, to bridge the gap between theory and practice, it is essential to supplement academic speculations with empirical insights from construction practitioners. A deep understanding of practical constraints, operational bottlenecks and on-the-ground requirements is vital to inform scalable and context-specific BIM–AI solutions. Additionally, identifying specific domains where BIM–AI integration can provide the greatest impact will support targeted implementation, improve return on investment and advance the industry’s broader digital transformation agenda.

BIM, while widely adopted for planning and design coordination, continues to present operational challenges within construction environments. One notable issue is the uneven utilization of BIM across professional roles, with BIM coordinators demonstrating the highest engagement and site-based personnel, such as superintendents, showing significantly lower usage levels (Britani and Thais, 2020; Hans et al., 2021). This disparity underscores a fundamental disconnect between digital planning tools and field-level application. The lack of structured BIM training programs exacerbates this issue, limiting the effective dissemination of BIM knowledge across diverse job roles and impeding its broader application on construction sites. Similarly, the implementation of AI in construction presents its own operational limitations. Studies by Charisma et al. (2020) and Kai and Nabil (2019) have highlighted several barriers to AI adoption, including the technical complexity of AI tools, insufficient intuitive design in authoring systems and the high resource demands associated with building Intelligent Instructional Systems (IIS) from scratch. Furthermore, the absence of interoperable and reusable AI components, coupled with limited empirical methods for evaluating AI performance, restricts scalability. Additional concerns such as data access limitations, communication challenges among stakeholders and skepticism regarding AI’s practical value further hinder implementation efforts.

While BIM and AI each face distinct operational hurdles, their integration introduces a new layer of complexity, rather than simply inheriting or resolving existing issues. Integrated BIM–AI systems, though rich with potential, bring forward a convergence of challenges from both domains – augmented by novel issues unique to their intersection. Khan et al. (2024) categorize these challenges into technological, organizational, managerial, data-centric, financial and knowledge-based barriers, reflecting the multifaceted nature of the integration process. From a technological standpoint, difficulties arise from the lack of interoperability between BIM and AI platforms, particularly in processing and extracting geometric and semantic data from BIM models. Li et al. (2024) and Du et al. (2024) note the absence of mature toolchains to convert Industry Foundation Classes (IFC) into AI-readable formats, posing a major bottleneck in automation and data-driven modeling. In addition, data standardization and quality assurance remain problematic, especially when dealing with fragmented or inconsistent datasets across construction phases.

Organizationally, the deficiency of cross-disciplinary expertise, as well as a limited understanding of AI among traditional construction professionals, inhibits full-scale adoption. Amanzadegan et al. (2024) emphasize the need for standardized training protocols, not only in the use of BIM and AI individually but also in their integrated application. Moreover, data protection concerns, especially in cloud-based and collaborative environments, introduce ethical and legal challenges that remain unresolved in many jurisdictions. Even in cases where the technical integration is feasible, managerial reluctance due to high upfront costs, uncertain ROI and lack of clear performance metrics undermines long-term commitment. Although studies like Ajirotutu et al. (2024) demonstrate that BIM–AI can enhance project efficiency, resource allocation and sustainability, these benefits remain largely theoretical or limited to controlled pilot projects. Real-world scalability is frequently impeded by the challenges outlined above.

Thus, the contribution of this body of research lies in identifying and categorizing the operational barriers that obstruct the effective integration of BIM and AI. Understanding these challenges is crucial for developing pragmatic solutions tailored to the construction sector. Rather than treating integration as a purely technological endeavor, it must be addressed as an organizational transformation requiring change management, workforce development and strategic investment in digital infrastructure. The documented discrepancies in BIM usage by these studies highlight a need for inclusive implementation strategies that consider diverse job roles and field realities. The analysis underscores the technical and systemic incompatibilities between BIM and AI, especially in data conversion and interoperability. Lack of training, resistance to change and financial risk aversion are shown to be critical non-technical impediments. A recurring theme is the absence of shared frameworks for data, training, ethics and evaluation – calling for industry-wide standards.

The evaluation of BIM performance in construction projects has traditionally relied on established capability and maturity models, such as the NBIMS Capability Maturity Model (CMM) and the BIM Maturity Matrix (BIM-MM). These frameworks provide structured guidelines to assess BIM competence across various levels, encompassing dimensions such as collaboration, data exchange, modeling proficiency and lifecycle integration (Asli and Onur, 2019). Such criteria not only aid in diagnosing an organization’s current BIM capabilities but also serve as strategic tools to identify areas for continuous improvement and innovation. Importantly, these models ensure that BIM applications are systematically embedded across all stages of a construction project – from design and preconstruction to execution and facility management. However, these BIM-focused evaluation models are domain-specific and largely static, often insufficient when extended to dynamic, data-driven systems like AI. AI systems require broader and more adaptive assessment frameworks. According to Reddy et al. (2021), AI evaluation extends beyond traditional performance metrics such as accuracy and precision, to include dimensions like scalability, usability, user adoption, fairness, interpretability and ethical reliability. These attributes are especially crucial for systems that influence human decision-making and operational outcomes, as is often the case in construction environments where AI is deployed to support predictive modeling, risk assessment or process automation.

Cihon (2019) further emphasizes that establishing well-defined AI assessment criteria is vital for fostering trust and accountability in AI deployment. Transparent evaluation mechanisms contribute to responsible AI usage, ensuring that systems are not only technically sound but also ethically and socially sustainable. Moreover, reporting and interpretability are increasingly viewed as non-negotiable features for advanced AI systems in sensitive and regulated industries such as construction and infrastructure. Despite the progress in the separate evaluation of BIM and AI, integrated BIM–AI systems lack a unified and standardized evaluation framework. This presents a significant barrier to their widespread adoption. Integrated BIM–AI applications, which combine digital modeling with AI-driven analytics or automation, introduce new operational paradigms that neither traditional BIM maturity models nor AI assessment standards can fully address in isolation. For instance, evaluating such systems requires metrics that simultaneously capture BIM’s data richness, interoperability and lifecycle value, and AI’s learning capacity, automation potential and ethical transparency.

Rane (2023) and Hugo et al. (2023) argue that the absence of holistic evaluation criteria impairs construction organizations’ ability to measure the true value and performance of integrated BIM–AI deployments. Without a clear benchmark, it becomes challenging to assess whether these systems enhance productivity, reduce risk or deliver long-term cost savings. Vicente et al. (2022) suggest that standardized evaluation criteria should encompass technical, organizational and human-centered dimensions, including integration quality, cross-platform compatibility, user interaction, adaptability to changing conditions and compliance with regulatory frameworks. This discourse highlights the incompatibility of existing BIM and AI evaluation standards when applied to integrated systems, underscoring the need for an interdisciplinary approach to assessment. Evaluation standards that reflect both technical capabilities and user-centered outcomes can support better alignment between technology deployment and business goals in construction. Integrating ethical evaluation criteria from AI research into BIM–AI systems ensures responsible implementation, a critical consideration in public infrastructure and urban development projects. A comprehensive evaluation framework would provide a benchmark for innovation, enabling organizations to experiment with confidence while measuring progress against defined performance indicators. While existing BIM and AI evaluation models offer foundational guidance, they are inadequate for assessing the synergistic and multifaceted nature of integrated BIM–AI systems. A comprehensive and adaptable evaluation framework is essential not only for validating performance but also for driving innovation, compliance and ethical implementation in the construction industry. Developing such a framework should be a research and policy priority to fully realize the transformative potential of BIM–AI integration.

This study adopts a dual-theoretical lens, drawing on the Strategic Alignment Model (SAM) and the Balanced Scorecard (BSC) framework to comprehensively conceptualize the integration of BIM with AI in construction operations. It examines the integration from three interconnected dimensions: (1) identifying potential domains for BIM–AI application, (2) diagnosing operational challenges that may hinder integration and (3) proposing a robust evaluation standard for integrated systems. This layered approach provides both a strategic alignment perspective (through SAM) and a performance evaluation perspective (through BSC), enabling a holistic understanding of how BIM–AI systems can be successfully deployed and sustained within construction organizations. The SAM serves as a foundational theory in information systems strategy. It emphasizes the need for strategic congruence between an organization’s business goals and its information technology (IT) capabilities. SAM identifies four critical domains: business strategy, IT strategy, organizational infrastructure and processes and IT infrastructure and processes. Alignment is achieved when these domains are coherently connected across two primary dimensions: strategic fit (between internal and external factors) and functional integration (across business and IT domains). In the context of BIM–AI integration, SAM is particularly valuable for illustrating how construction firms must harmonize technological innovation (AI capabilities within BIM environments) with organizational intent (project performance, stakeholder engagement and resource optimization). The model thus ensures that the adoption of BIM–AI is not isolated to the IT department but embedded into the broader strategic and operational fabric of the organization.

Empirical research by Yolande and Blaize (2007) has validated SAM’s utility in guiding digital transformation efforts, especially in complex, multi-stakeholder industries such as construction. By applying SAM, this study identifies eight core domains for integrating BIM with AI, such as automated progress tracking, intelligent design validation, risk forecasting and sustainability modeling. These domains are mapped against the organization’s strategic vision, IT readiness and operational capacities to ensure alignment and maximize impact. Complementing SAM, the BSC framework introduces a performance-based strategic management approach that translates high-level goals into measurable actions. BSC views organizational success through four interrelated perspectives: financial performance, customer satisfaction, internal business processes and learning and growth. When applied to BIM–AI integration, the BSC framework enables construction firms to assess how integrated technologies impact not just immediate operational metrics, but also long-term capability development, innovation capacity and stakeholder value delivery. For instance, under the financial perspective, firms can evaluate cost savings from reduced rework and optimized resource usage through AI-enhanced BIM. The customer perspective can track improvements in client satisfaction due to faster project delivery and higher transparency. The internal process perspective assesses gains in process efficiency and quality control through automation and predictive analytics. Finally, the learning and growth perspective considers how BIM–AI systems foster organizational knowledge, training and cross-disciplinary collaboration.

The integration of SAM and BSC in this study is both conceptual and diagnostic. It not only frames the discussion around where and how BIM–AI systems should be applied but also provides criteria to evaluate their strategic and operational efficacy. Specifically, the research identifies seven key operational challenges – including data interoperability, training deficits, cost barriers and ethical concerns – that must be mitigated to enable successful integration. In parallel, it proposes seven evaluative criteria drawn from both SAM and BSC perspectives to ensure continuous performance monitoring, strategic alignment and user satisfaction. Figure 1 presents a visual synthesis of this integrated framework. It maps the eight identified BIM–AI application domains against the operational challenges and evaluation criteria, providing a comprehensive roadmap for construction firms seeking to adopt and scale BIM–AI systems. By leveraging these dual frameworks, the study contributes a nuanced, systems-level understanding of BIM–AI integration – one that balances technical innovation with strategic coherence and performance accountability.

This study employs a methodologically rigorous design grounded in the application of multiple research paradigms – including positivism, realism, critical theory and interpretivism – to ensure philosophical balance, reduce methodological bias and enhance the quality of research findings. By drawing from these diverse paradigms, the study improves the reliability, validity and generalizability of its outcomes across different epistemological domains and disciplinary perspectives. Positivism serves as the core philosophical foundation of this research, favoring quantitative, objective evidence to investigate the relationships among latent variables in the integration of BIM and AI. It aligns with the study’s reliance on survey data, statistical modeling and confirmatory factor analysis (CFA), aiming to identify generalizable patterns through observable and measurable data. Realism, while sharing positivism’s emphasis on empirical rigor, introduces a context-sensitive lens by acknowledging that human perceptions and experiences shape technological adoption. This is particularly relevant in construction environments, where organizational culture and role-specific dynamics influence the use of BIM and AI tools. Interpretivism and critical theory offer complementary perspectives. Interpretivism highlights the role of subjective meaning-making among construction professionals, helping researchers to contextualize the operational challenges reported in the data. Critical theory enables reflection on power dynamics, access disparities and institutional inertia, all of which affect the successful adoption of BIM–AI technologies. These philosophical orientations indirectly informed the study’s problem framing and justification for a broader, inclusive sample.

To ensure theoretical depth and multidimensional analysis, the research employed a theoretical triangulation strategy, drawing from two well-established models: the SAM and the BSC. The SAM guided the development of constructs related to strategic integration and technological alignment, while the BSC informed the operational performance and evaluation metrics. Together, these frameworks supported the design of a structured and theory-driven questionnaire that addressed three primary areas: (1) application domains of BIM–AI integration, (2) operational challenges and (3) assessment criteria. Although the overarching approach is positivist, the study employed within-method triangulation, meaning the quantitative strategy was reinforced by using multiple theoretical lenses to analyze the same dataset. This strengthened the construct validity of the measurement models and minimized confirmation bias. Data were collected via a survey-based approach from the full membership of the Federation of Construction Industry (FOCI) in Nigeria – an authoritative body overseeing infrastructure and construction projects. Three representatives were surveyed per company, resulting in the dissemination of 240 questionnaires, of which 214 fully completed responses were used in the final analysis (a high response rate of 89.2%).

The questionnaire was logically structured to optimize clarity and flow: section 1 collected demographic and professional profile data. Section 2 focused on potential areas for AI–BIM synergy in operational processes. Section 3 addressed operational barriers to integration. Section 4 evaluated criteria for assessing the effectiveness of integrated BIM–AI systems. These variables were explicitly derived from the conceptual foundations of SAM and BSC to ensure coherence between theory and measurement. To further enhance data integrity, a mixed-mode data-collection strategy was used – combining both digital and paper-based surveys. This approach reduced non-response bias, expanded accessibility and minimized measurement error associated with a single-mode delivery. It also enabled greater reach across organizational hierarchies and geographical locations. Respondent profiling confirmed a highly competent participant base in terms of educational attainment (32.5% held B.Sc. degrees, 31% had OND qualifications, 20% had HNDs, 15% held M.Sc. degrees and 1.5% had PhDs), professional roles (42.9% were project managers, 25% construction managers, 17.3% site managers, 10% directors and 4.8% procurement managers), years of experience (46% had over 21 years in the industry, and another 40.7% had between 6 and 20 years, indicating substantial domain expertise) and professional disciplines (33.1% were civil engineers, 27.2% building engineers, 24.7% quantity surveyors, 8.9% surveyors and 6.1% architectural designers).

The analytical phase employed CFA to validate the measurement constructs and assess the fit and consistency of observed variables with the theoretical framework. CFA was executed in six stages: (1) specification of latent variables, (2) conceptual mapping of constructs (applications, challenges, evaluation), (3) model identification, (4) data collection, (5) estimation of model parameters and (6) validation and interpretation. This method allowed the study to test theory-driven hypotheses about the underlying structure of BIM–AI integration and assess how well real-world data aligned with theoretical expectations. The CFA results substantiated the robustness of the sub-constructs derived from SAM and BSC, offering a reliable empirical basis for identifying the most critical factors affecting BIM–AI adoption. The study’s findings offer a conceptual scaffold for future research and a practical roadmap for construction firms seeking to implement integrated BIM–AI systems. The insights derived from this study are not merely descriptive – they serve as foundational evidence for policy recommendations, training program development and strategic planning within the construction sector.

In order to explore the potential integration of BIM and AI in construction organizations’ operations, the theoretical framework identified nine domains, which were assessed using 54 factors. The participants were requested to express their degree of agreement with the factors using a five-point Likert scale. The data obtained from the participants were analyzed using confirmatory factor analysis. The findings of the confirmatory factor analysis are presented in Table 1. The findings indicate that all the hypothesized domains of integrated BIM–AI are dependable, substantial and elucidate the fusion of BIM with AI. The Cronbach Alpha, Eigen value and average variance explained for the integration of BIM with AI exceeded 0.80, 0.00 and 0.50, respectively. The relevance of the variables in integrating BIM with AI was assessed based on their factor loading. A factor loading of 0.7 or above indicates a strong level of significance. Table 1 demonstrates that among the eleven variables considered for integrated BIM–AI in smart construction management, nine variables (automated design and rule checking, 3D as-built reconstruction, building performance analysis, balance of the construction team, automated progress monitoring, automated safety management, automated quality management, automated contract management and automated site progress monitoring) exhibited strong significance, with factor loadings of 0.70 and above. Event log mining, with a factor loading of 0.62, and digital twins of construction site items, with a factor loading of 0.52, were determined to be acceptable.

Regarding integrated BIM–AI for waste management, the respondents stated that it may be highly valuable for waste quantitative evaluation (F_L = 0.86) and generating sustainable waste disposal methods (F_L = 0.73). The respondents agreed that integrated BIM–AI can be significantly applied for many different uses in facility management. These include 3D as-built documentation (F_L = 0.77), 3D scan-to-BIM reconstruction (F_L = 0.81), predictive property valuation (F_L = 0.83), automatic property data classification (F_L = 0.84), predictive energy management (F_L = 0.92), thermal image analysis for defects (F_L = 0.71), automatic recognition of architectural elements (F_L = 0.70), automatic historic building management (F_L = 0.80) and BIM–AI-driven land supply management (F_L = 0.80). The respondents responded that combined BIM and AI can be very applicable for data management. There is also an intuitive voice assistant for managing BIM data, with a factor loading of 0.84. The BIM content recommender system powered by AI has a factor loading of 0.82. Additionally, there is an AI voice assistant interface that can translate verbal requests with a factor loading of 0.86. Lastly, there is a centralized data-management system with a factor loading of 0.78. The survey participants also concurred that the combination of BIM and AI can be effectively utilized for identifying defects and damages in the following domains: automated assessment of seismic damages in buildings (F_L = 0.80), replication of defects in digital form (F_L = 0.70), automated analysis of damage caused by hazards (F_L = 0.73) and AI models for predicting defects (F_L = 0.71).

The respondents have agreed that the integrated BIM–AI application areas for sustainable design and modeling are BIM–AI-enabled energy-efficient building design (F_L = 0.92), BIM–AI-based sustainable design decision-making (F_L = 0.93), predictive AI models for HVAC energy consumption (F_L = 0.88) and intelligent utilization of construction materials (F_L = 0.77). The areas of application for integrated BIM–AI in project management, as indicated by the respondents, include a BIM–AI-based system for evaluating project performance (F_L = 0.71), modeling functions and processes using BIM–AI (F_L = 0.70), automating project scheduling (F_L = 0.90), generating and exploring creative project designs (F_L = 0.90), collaborating on project design (F_L = 0.71), authoring and reviewing BIM models automatically (F_L = 0.77), estimating project costs automatically (F_L = 0.81), finding paths automatically (F_L = 0.66), checking code compliance automatically (F_L = 0.71), detecting clashes automatically (F_L = 0.74) and simulating construction automatically (F_L = 0.86). The key areas where integrated BIM–AI can be applied in construction technology are as follows: intelligent recognition algorithms for monitoring structural damage (F_L = 0.76), predictive AI models for forecasting structural damage (F_L = 0.77), AI-generated structural engineering BIM models (F_L = 0.76) and AI-generated mechanical and electrical engineering BIM models (F_L = 0.74).

The key operational challenges of integrated BIM–AI in construction organizations were conceptualized based on the theoretical insights from the SAM and BS, as explained in Figure 1. These challenges include issues arising from the interoperability of data between the BIM and AI platforms, employees’ growth and development, governance for integrated BIM–AI, infrastructure for integrated BIM–AI, finances for integrated BIM–AI, internal processes for integrated BIM–AI and client relations through the BIM–AI platform. Based on the confirmatory factor analysis results of the respondents’ replies, as shown in Table 2, all the sub-constructs are dependable and effectively illustrate the operational constraints of integrating BIM–AI in construction organizations. The Cronbach Alpha, Eigen value and average variance explained for the sub-constructs exceeded 0.80, 0.00 and 0.50, respectively.

The factor loading analysis of the operational challenges arising from the interoperability of data between BIM and AI platforms showed that the following challenges are significant: data security and data training (F_L = 0.77), differences in AI and BIM models (F_L = 0.86), lack of information in existing BIM software to support AI features (F_L = 0.88) and unfamiliarity with user interfaces and AI features in BIM software (F_L = 0.78). These challenges are expected to be encountered when integrating BIM and AI. The findings also indicated that the learning culture (F_L = 0.78) and learning style (F_L = 0.89) of the personnel in construction organizations pose operational hurdles for their growth and development in the context of integrated BIM–AI. Two major operational issues that will impact workers’ growth and development with regard to the implementation of integrated BIM–AI in construction organizations are brain drain (F_L = 0.96) and employees’ motivation (F_L = 0.88). The respondents have identified several significant operational challenges that will impact governance for integrated BIM–AI in construction organizations. These challenges include corruption in training and investment (F_L = 0.87), the need to strengthen the decision-making process (F_L = 0.71) and ensuring commitment to investment in BIM–AI platforms (F_L = 0.77).

The main operational difficulties related to the infrastructure for integrated BIM–AI in construction organizations consist of insufficient infrastructure (F_L = 0.71), unreliability of infrastructure (F_L = 0.72) and limited availability of software that integrates BIM and AI (F_L = 0.81). The survey participants concurred that the implementation of integrated BIM–AI in construction organizations will face the following obstacles: insufficient cash flow and debt (F_L = 0.70), unstable funding (F_L = 0.62), budget limitations (F_L = 0.81), insufficiency of working capital (F_L = 0.61) and economic downturns (F_L = 0.76). The primary operational problems related to the internal processing of integrated BIM–AI in construction organizations are inefficient workflows (F_L = 0.70), inadequate support for innovation (F_L = 0.73), lack of collaboration (F_L = 0.66) and a negative culture (F_L = 0.71). The respondents have identified several key operational challenges associated with client relations through the BIM–AI platform. These challenges include miscommunication and communication gaps (F_L = 0.68), differing expectations (F_L = 0.70), uncertainty about the output of integrated BIM–AI (F_L = 0.86) and work overload (F_L = 0.80).

According to the research framework depicted in Figure 1, this study conceptualized the fact that the assessment criteria for integrated BIM–AI in construction organizations should be in line with the operational aspects of integrated BIM–AI. The evaluation standards for integrated BIM–AI in construction organizations were assessed based on the interoperability of data between BIM and AI platforms, employee growth and development, governance for integrated BIM–AI, infrastructure for integrated BIM–AI, finances for integrated BIM–AI, internal processes for integrated BIM–AI and client relations through the BIM–AI platform. The confirmatory factor analysis findings, as shown in Table 3, indicate that all the sub-constructs are dependable and elucidate the assessment criteria for integrated BIM–AI in construction organizations. The Cronbach Alpha, Eigen value and average variance explained for the sub-constructs exceeded 0.80, 0.00 and 0.50, respectively.

The factor loading for the measured variables used as evaluation standards for interoperability of data from BIM and AI platforms, as presented in Table 3, indicated that data exchange (F_L = 0.70), data synchronization (F_L = 0.70), data security (F_L = 0.81) and data privacy (F_L = 0.74) are the important evaluation criteria for interoperability of integrated BIM and AI. The key criteria for assessing the progress and advancement of personnel through the implementation of integrated BIM–AI are job satisfaction (F_L = 0.90), employee turnover (F_L = 0.71), specialized knowledge and skills (F_L = 0.81) and training opportunities (F_L = 0.77). The participants reported that transparency (F_L = 0.77), communication (F_L = 0.71), innovation (F_L = 0.80) and productivity (F_L = 0.78) are important factors to consider while evaluating the governance of integrated BIM–AI. The findings presented in Table 3 indicate that the assessment criteria for integrated BIM–AI infrastructure should encompass processing accuracy (F_L = 0.86), prediction accuracy (F_L = 0.81), visualization quality (F_L = 0.82), ease of use (F_L = 0.78) and operating efficiency (F_L = 0.86).

The financial analysis of integrated BIM–AI may be assessed using return on investment (F_L = 0.78), cash flow (F_L = 0.71) and financial results (F_L = 0.92), as shown in Table 3. The criteria for assessing the internal process of integrated BIM–AI, based on the findings, consist of activities per function (F_L = 0.71), automation (F_L = 0.72), process alignment (F_L = 0.71) and response time (F_L = 0.84). As per the agreement of the respondents, the evaluation of client relations through the BIM–AI platform may be done using the following criteria: client satisfaction rate (F_L = 0.88), client retention (F_L = 0.96), quality performance (F_L = 0.90), cost reduction (F_L = 0.90) and client acquisition (F_L = 0.88).

The findings of this study offer compelling empirical support for the integration of BIM and AI in construction operations across a wide range of functional domains. The results from CFA affirm that the nine hypothesized domains of BIM–AI integration are statistically robust and conceptually coherent. The high factor loadings across most indicators underscore the strong internal consistency of the domains. This validates the theoretical proposition that BIM–AI fusion is not only technically feasible but also operationally impactful across various construction functions. This domain revealed very strong support, with nine out of eleven variables demonstrating high factor loadings (≥0.70). This suggests that stakeholders recognize AI-enhanced BIM systems as transformative tools for automating routine tasks (progress monitoring, safety and quality management), improving design accuracy (automated rule-checking, 3D as-built reconstruction) and facilitating real-time site updates. While event log mining and digital twins showed slightly lower scores, they still fall within acceptable thresholds, indicating emerging relevance in practice that may grow as digital maturity improves.

With high factor loadings for waste quantification and sustainable disposal method generation, this domain reflects a growing awareness of environmental accountability in construction. Respondents perceive BIM–AI as a critical enabler of intelligent material tracking and reduction of environmental impact, aligning with circular economy principles. The results highlight BIM–AI’s powerful capabilities in predictive analytics for property valuation and energy management, automatic documentation and classification of facility assets and enhanced visual inspection and fault detection (thermal imaging, architectural recognition). This supports the notion that facility lifecycle management stands to gain substantially from BIM–AI integration due to its reliance on both structured data and visual/semantic processing. The results indicate strong user confidence in AI-powered BIM navigation (voice assistants), content recommendation systems and centralized data governance. These results emphasize the increasing user-centric shift in construction data systems – toward systems that are intuitive, voice-enabled and automated, reducing friction in digital workflows.

BIM–AI is perceived as highly effective for post-disaster assessment, predictive maintenance and real-time defect identification. This supports the integration of BIM–AI into risk management and resilience frameworks, especially in areas prone to natural hazards or complex infrastructure challenges. The results reinforce the paradigm shift toward green and smart buildings, where BIM–AI tools are instrumental in lifecycle carbon modeling, intelligent materials selection and HVAC system optimization. This demonstrates strategic value, as sustainability has become a core concern for both regulators and clients. The results reflect wide acceptance of BIM-AI’s role in automating scheduling and costing, enhancing collaborative design and ensuring code compliance and clash detection. The results also indicate the growing trust in BIM–AI systems for structural monitoring, AI-generated engineering models and forecasting structural health. This suggests that AI-enhanced BIM can move beyond static modeling to become an active decision-support tool in structural engineering and MEP (mechanical, electrical, plumbing) design.

These findings validate that construction professionals regard integrated BIM-AI systems as critical enablers of productivity, safety, sustainability and decision-making. Domains with the highest impact perception include sustainability modeling, facility management and data management, suggesting where the most immediate value may lie. The findings point to digital twins and event log mining as emerging technologies whose potential may be unlocked with further training, infrastructure and innovation. The research highlights a technological convergence where AI serves as the intelligence layer over BIM’s structured digital twin environment – unlocking new efficiencies, analytics and automation across the construction value chain. The findings from this current study corroborate and expand upon the existing body of literature on BIM-AI integration in construction operations. Samad et al. (2022) identified the lack of automation and non-value-adding tasks as key drawbacks of traditional BIM tools. This study validates this claim empirically: construction professionals confirmed that integrating AI with BIM significantly enhances automation in tasks like progress monitoring, rule checking, safety, quality and contract management – resolving many of the inefficiencies that previously limited BIM’s impact.

Studies like Tan et al. (2022) and Sofiat et al. (2022) emphasized productivity, cost control and safety as major application areas. The findings of this study support these observations, especially in the domains of project management and smart construction, where respondents strongly endorsed AI–BIM for automated scheduling, cost estimation and safety management. Scholars including Heidari et al. (2024) and Ghasemi Poor Sabet and Heap-Yih (2020) noted the role of BIM–AI integration in driving digital transformation. This study confirms that practitioners perceive AI as a critical intelligence layer that extends BIM’s capabilities – from static modeling to dynamic, predictive and automated systems. Albert et al. (2022), Rangasamy and Yang (2024) and Alavi et al. (2024) explored integration with IoT, blockchain and digital twins. The findings of this study include positive though moderate factor loadings for digital twins and event log mining, suggesting early but growing acceptance – thus empirically validating these speculative propositions. Saleh et al. (2024) advocated for the fusion of BIM–AI to support renewable energy, sustainable materials and intelligent controls. This current study not only supports this but demonstrates strongest factor loadings in the sustainability domain (AI-based sustainable design decision-making), indicating widespread professional consensus and operational maturity in this area.

This current study also adds novel insight. Previous studies often theorized potential applications of BIM–AI in broad terms. This study uniquely contributes by offering quantitative evidence across nine specific domains with 54 tested variables – demonstrating which use cases practitioners deem most valuable, and at what significance level. While prior literature focused more on construction processes and site operations, your findings highlight strong industry support for AI–BIM in facility management (predictive energy management) and data management (intuitive voice assistant), areas that have received comparatively less empirical attention in earlier studies. Much of the past literature was speculative or based on conceptual models. This research fills this gap by capturing the practical insights and preferences of industry professionals, offering a grounded understanding of how integration is viewed on the ground – a point emphasized as crucial in your literature summary.

The findings of this study provide a comprehensive, empirically grounded view of the operational challenges faced by construction organizations when integrating BIM with AI. The results of this study demonstrate that the theoretical operational dimensions (interoperability, workforce, governance, infrastructure, finances, internal processes and client relations) are not only conceptually valid but also resonate with practitioners’ experiences. Findings on interoperability of BIM–AI Platforms such as lack of AI-supportive data structures in current BIM tools, incompatibility of AI and BIM models, data security and training and unfamiliar interfaces suggest a technical misalignment between existing BIM systems and emerging AI tools, creating friction in smooth integration and usage. Key issues regarding employees’ growth and development, such as brain drain, low motivation, mismatched learning styles and inadequate learning culture, represent serious human capital constraints – particularly the very high loading for brain drain points to fears that the most capable personnel might exit the sector, undermining innovation and continuity.

The findings revealed that the key operational challenges in governance of BIM–AI Implementation are corruption in training/investment, weak decision-making and lack of investment commitment. This indicates that organizational governance and ethical issues could critically hinder implementation, especially in contexts with weak institutional oversight or fragmented leadership structures. Infrastructure readiness’ key operational challenges are limited integrated BIM–AI software, unreliability of infrastructure and insufficient base infrastructure. These suggest foundational technological barriers, particularly in environments where digital infrastructure is still underdeveloped or not standardized. The key challenges in financial constraints include budget limitations, economic downturns, poor cash flow, unstable funding and inadequate working capital. These challenges reflect financial fragility – organizations may understand the value of BIM–AI but lack the liquidity, investment stability or resilience to fund adoption and maintenance. Internal processes’ major challenges inadequate innovation support, negative culture, inefficient workflows and lack of collaboration. These indicate organizational inertia and fragmentation, where even with the right tools, internal misalignment and cultural resistance can delay or derail integration. Uncertainty about BIM–AI outputs, work overload, miscommunication and differing expectations were the major challenges in client relations via BIM–AI. This highlights the external-facing risks of integration – clients may mistrust or misunderstand AI-generated outcomes, and unclear expectations may erode satisfaction or trust.

High significance in several domains – like brain drain, software limitations and uncertain outputs – indicate that these are priority issues needing targeted mitigation strategies. There is a notable blend of technical, human, financial and organizational challenges, underscoring that BIM–AI integration is not just a technology issue, but a strategic transformation initiative. These findings also suggest a systemic interdependence: for instance, poor infrastructure exacerbates workflow inefficiencies, while financial constraints limit governance and training capacity. The findings from this research corroborate and build upon the previous studies in several key ways, while also offering distinct contributions through its empirical depth and practitioner-centered validation. This study provides a strong evidence of interoperability problems between BIM and AI (e.g., incompatible data models, insufficient BIM software support for AI and user interface unfamiliarity). Du et al. (2024), Li et al. (2024) and Amanzadegan et al. (2024) highlighted similar problems, such as lack of interoperability standards, geometric data extraction issues and conversion limitations with IFC formats. The findings of this study strongly reinforce earlier academic predictions by validating them with practitioner insights. This shows that these issues are not just speculative but actively impact organizations in the field.

Brain drain, low motivation, incompatible learning styles and lack of a learning culture were found as major constraints in this study. Lack of familiarity with AI–BIM tools and interfaces was also a challenge, as found in this study. Previous studies by Britani and Thais (2020) and Hans et al. (2021) reported a lack of BIM training, especially among site staff like superintendents. Amanzadegan et al. (2024) noted inadequate expert training as a barrier. This research empirically confirms these earlier observations and deepens them by identifying specific psychological and organizational factors like motivation and learning style mismatches, which are rarely quantified in earlier work. As found in this study, governance issues include corruption in training investments, weak decision-making and lack of commitment to BIM–AI investments. Internal inefficiencies include negative culture, lack of collaboration and poor support for innovation. Khan et al. (2024) emphasized organizational and managerial barriers as key obstacles. Kai and Nabil (2019) mentioned a disconnect between users and AI tools due to unintelligent systems and poor authoring environments. While prior studies noted general managerial constraints, this study’s findings go further by detailing the institutional weaknesses (e.g., corruption) and cultural roadblocks, offering a clearer picture of internal organizational dysfunction.

This study found financial instability, budget limitations and economic downturns to be significant barriers to BIM–AI adoption. Khan et al. (2024) and Ajirotutu et al. (2024) also reported financial and cost-related issues, such as high implementation costs. There is strong alignment here. This study, however, quantifies these issues through factor loadings and distinguishes between types of financial stress (e.g., debt vs. unstable funding), offering more granular insight. This study reported lack of data training, absence of necessary BIM information to support AI, information gaps and uncertainties about AI outputs among clients as data and knowledge management challenges. Charisma et al. (2020) and Kai and Nabil (2019) discussed poor authoring tools, lack of reusable AI components and difficulty assessing AI application effectiveness. Du et al. (2024) stressed data conversion challenges and poor toolchains. This study validates and expands these theoretical concerns by confirming that data knowledge, training and uncertainty are operational pain points, especially in client interactions. This study found that clients face communication gaps, unclear expectations, work overload and uncertainty about AI–BIM outputs. Few prior studies directly addressed client-facing operational issues, though Ajirotutu et al. (2024) touched on stakeholder collaboration benefits. This current research brings new attention to external stakeholder dynamics, which were underexplored in prior literature, highlighting the need for clearer communication and expectation management in client–BIM–AI interactions.

While prior studies focused largely on theoretical or qualitative insights, this study quantitatively validates these challenges using factor loadings and reliability metrics. The integration of SAM and BS, as done in this study provides a holistic operational view, connecting strategy, infrastructure, personnel and external factors in a single conceptual model. Unlike most previous studies, this research emphasizes the practical voices of industry practitioners, giving a real-world grounding to the issues discussed in literature. This study’s findings corroborate much of the existing literature on the challenges of BIM and AI integration but also extend and enrich it through empirical evidence, greater conceptual depth and a broader operational lens. They offer actionable insights that bridge the gap between academic theory and field-level implementation, providing a strong foundation for both further research and informed decision-making in the construction industry.

The findings from this study provide a comprehensive and empirically grounded framework for evaluating the performance and implementation success of integrated BIM–AI systems in construction organizations. These evaluation criteria – conceptualized across seven core operational domains – offer a multi-dimensional assessment tool that aligns performance measurement with the operational realities of integrating BIM and AI technologies. The CFA results confirm that all seven sub-constructs are statistically robust and reliable for assessing BIM–AI integration in practice. The findings emphasize that data interoperability is not just a technical concern but a performance dimension. Data security and privacy reflect industry concerns about cybersecurity and confidentiality in AI-integrated BIM environments.

Findings on employee growth and development reflect the increasing recognition that human capital metrics are central to technology integration success. The high factor loading for job satisfaction suggests that employee well-being is a sensitive and reliable proxy for implementation success, potentially affecting retention and productivity. The results also underscore the need for strong internal governance structures. Innovation and productivity stand out as both goals and evaluative benchmarks, suggesting that integrated BIM–AI must be embedded within an innovation-friendly governance culture. Findings on infrastructure assessment criteria reflect a performance-oriented infrastructure evaluation. High loadings for operating and processing accuracy suggest that technical precision and system usability are critical for user adoption and efficiency. The standout loading for financial results indicates that stakeholders prioritize clear financial outcomes when evaluating BIM–AI success. This aligns with the broader push for evidence-based return on tech investments in the construction sector. The results on internal processes evaluation suggest a shift toward agility and process efficiency as benchmarks of successful integration. High response time loading signals that real-time decision-making and adaptive processes are highly valued.

Client relations through BIM–AI was found to have the highest factor loadings overall, indicating that client-facing outcomes are the most sensitive and measurable indicators of BIM–AI success. The emphasis on retention and satisfaction highlights how technology adoption is evaluated through the lens of end-user value. This study offers a strategic evaluation model that goes beyond technical performance by integrating human resource development, governance innovation, financial accountability, operational responsiveness and client value delivery. Each domain is supported by strong empirical data, making this framework suitable for real-world performance audits, strategic planning and risk mitigation in BIM–AI deployment. This positions the framework as a practical tool for construction firms seeking to implement or scale BIM–AI integration, while also contributing significantly to academic discourse on technology evaluation in the built environment. The insights from the study corroborate and advance previous findings on BIM and AI assessment frameworks by offering a unified, empirically validated evaluation structure specifically tailored for integrated BIM–AI systems – an area that earlier research has largely acknowledged as lacking clear, comprehensive criteria. This study aligns with established BIM Capability and Maturity Models by Asli and Onur (2019) and BIM evaluation models (NBIMS CMM, BIM-MM) by recognizing multiple operational dimensions of BIM evaluation such as infrastructure, process alignment and employee competency; and highlighting the importance of skill development, system usability and performance metrics – themes central to BIM maturity assessments. The framework from this study expands this by integrating AI-specific dimensions (e.g., prediction accuracy, data privacy, algorithmic utility), thus addressing the gap in existing BIM-centric models when applied to hybrid BIM–AI systems.

This study also resonates with the AI literature by Reddy et al. (2021) and Cihon (2019) that emphasizes performance beyond accuracy (utility, adoption, ethical considerations). System trust and transparency, as seen in governance evaluation criteria like transparency and communication; as well as usability and interpretability, as found in this study, reflected in ease of use and visualization quality. Where prior studies consider AI assessment largely in isolation, the integrated approach adopted in this study evaluates AI within the construction context, embedding AI assessment into organizational, financial and client-centric metrics that are often neglected in general AI frameworks. Rane (2023), Hugo et al. (2023) and Vicente et al. (2022) point out that no standard evaluation framework exists for integrated BIM–AI, making assessment fragmented and insufficient. This study directly addresses this gap by creating an integrated, multi-domain assessment framework with empirical backing; proposing a synthesized model where BIM and AI evaluation dimensions are not only merged but contextually grounded in construction practice and enabling cross-domain evaluation – for example, linking data interoperability to employee development or infrastructure performance to client satisfaction – thus promoting holistic project evaluation. This fulfills the call by Vicente et al. (2022) for criteria that can help assess BIM–AI effectiveness and identify critical success factors. This study is unique as it presents a unified and integrated evaluation for BIM–AI; while previous studies present separate evaluation models for BIM and AI. The study corroborates the idea that assessment must go beyond technical accuracy or maturity levels. It also advances the literature by presenting integrated evaluation criteria grounded in operational realities – thus serving as a blueprint for both implementation and performance auditing of BIM–AI systems.

This study provides strong empirical support for the conceptual coherence and operational relevance of nine BIM–AI integration domains. Areas like sustainability, facility management and data governance stood out as the most impactful. BIM–AI is recognized not just as a technical merger but as a transformative enabler of automation, real-time decision-making and operational efficiency across construction processes – ranging from design to facility management. Digital twins and event log mining, though not as mature, are gaining recognition and are expected to become central as digital maturity improves. A new integrated multi-dimensional framework was developed, offering statistically validated assessment tools across seven operational dimensions, including governance, financial resilience, client relations and infrastructure usability. Integration is hampered by a blend of technical (interoperability, software limitations), human (brain drain, low motivation), financial (budget instability) and organizational (poor governance, negative culture) barriers – showing the need for holistic strategies beyond technical implementation. Among all dimensions, end-user experience and satisfaction are the most sensitive indicators of integration success.

This study concludes that BIM–AI integration has moved from a conceptual possibility to an operational imperative. It can drive productivity, sustainability, safety and predictive capacity in construction. Success hinges on addressing foundational challenges in workforce development, infrastructure, governance and client engagement. The integration should be approached as a transformation initiative, not just a tech upgrade. The proposed framework offers a pioneering method to evaluate BIM–AI systems holistically, grounded in empirical data and operational realities. This study provides one of the most comprehensive quantitative validations of BIM–AI integration, covering 54 variables across 9 domains, thus filling the gap in empirical literature dominated by conceptual models. The study presents the first integrated, multi-domain, performance-oriented framework for assessing BIM–AI implementation, combining strategic, technical, organizational and human metrics. By quantifying challenges such as brain drain, financial fragility and client mistrust, the study adds new depth to known obstacles, including psychological and ethical dimensions not thoroughly addressed in prior studies. Unlike previous literature that leaned heavily on academic theorizing, this study captures real-world practitioner insights – validating and refining existing models from a field-centric viewpoint.

The multi-dimensional framework can serve as a basis for longitudinal studies, comparative analysis across regions and sector-specific BIM–AI evaluations. Integration challenges span technology, psychology, finance and governance, indicating the need for cross-disciplinary research in construction informatics, HR and organizational behavior. The evaluation framework can guide implementation audits, readiness assessments and investment decisions in construction organizations. Training, motivation and learning culture must be prioritized alongside software and infrastructure. Systems must focus on usability, communication and trust-building with clients – critical to achieving long-term value and adoption. Future studies should track the implementation of BIM–AI systems over time using the proposed model to observe long-term outcomes and system evolution. As emerging but underutilized tools, digital twins and event log mining require further research on their value realization, training needs and integration models. How organizational culture, employee psychology and change management affect BIM–AI adoption and usage effectiveness should be investigated. Models to bridge communication gaps between technical teams and clients, focusing on interpretability, expectations and trust in AI-generated outputs should be developed. Further studies on data privacy, transparency, algorithmic accountability and how to structure ethical governance in BIM–AI environments are also required.