Mapping Cost Intersection Through LCC, BIM, and AI: A Systematic Literature Review for Future Opportunities

Abstract

The increasing integration of digital technologies in the construction sector is transforming the processes of buildings design, management, and evaluation throughout their life cycle. Life Cycle Costing (LCC), Building Information Modeling (BIM), and openBIM standards play a key role in promoting economic and environmental sustainability. More recently, Artificial Intelligence (AI) has unlocked novel possibilities for data-driven decision-making and cost optimization. However, the integration of LCC, BIM, and AI is insufficiently explored in the current literature. This study presents a systematic literature review (SLR) aimed at analyzing two distinct lines of research, LCC–BIM and LCC–AI, and identifying underexplored opportunities for their future convergence. A dual-stream approach was adopted to analyze scientific contributions based on LCC–BIM and LCC–AI separately, using bibliometric analysis and the systematic screening of peer-reviewed articles from 2015 to 2025. The findings reveal that while LCC–BIM integration shows growing methodological maturity, AI-based applications are still in an early stage, with limited implementation in construction-specific contexts. The review identifies key challenges, including data fragmentation, a lack of interoperability, and limited standardization, as significant impediments to integrated digital workflows. By highlighting these gaps and proposing actionable future directions, the paper outlines future research directions focused on open data models, AI-enhanced cost estimation, and the development of interoperable frameworks to support sustainable and intelligent cost management in the Architecture, Engineering, and Construction (AEC) sector.

1. Introduction

In recent years, the integration of Life-Cycle Cost (LCC) methods and Building Information Modeling (BIM) has garnered increasing attention in the construction sector, establishing a novel framework for the sustainable and data-driven management of built assets across their entire life cycle. The growing demand for tools capable of supporting informed decision-making, based on structured, interoperable, and updatable information has driven research towards solutions that combine digital information modeling with advanced techniques of predictive analytics and process automation.

Today, BIM is a fundamental enabler of the digital transformation of the Architecture, Engineering, and Construction (AEC) sector, offering a multidimensional and integrated representation of buildings. Within this context, its integration with LCC facilitates early-stage, transparent, and comprehensive cost evaluations, overcoming traditional fragmented approaches [1,2,3,4,5,6]. Several studies have shown that BIM-based LCC can enhance the accuracy, auditability, and automation of cost estimation, while also incorporating environmental and social dimensions to support sustainable decision-making [5,6].

However, several challenges persist. One critical issue concerns the structuring of cost-related information within BIM models, especially in openBIM contexts, where the absence of appropriate classes, attributes, and properties limits full interoperability [3,7]. Additionally, external factors such as cost heterogeneity and obsolescence [8,9], variability in asset service life [8], and uncertainties in discount rates for Net Present Value (NPV) calculations [9,10,11] continue to affect the robustness of BIM-based LCC assessments [3,4].

While LCC–BIM integration has been widely investigated and is advancing in methodological maturity, several limitations still hinder its adoption at scale. These include interoperability gaps, data fragmentation, and a lack of standardized cost attributes within openBIM environments.

In parallel, the emergence of Artificial Intelligence (AI), especially through Machine Learning (ML) and Neural Networks (NNs), is opening new avenues for improving cost management. Several studies demonstrate the potential of AI to enhance cost prediction, uncertainty modeling, and decision-making, leveraging historical datasets to foster data-driven processes [6,11,12].

However, the widespread and systematic adoption of AI in LCC workflows remains in an early stage [13], constrained by the limited availability of large, high-quality, and structured datasets based on building component costs [11]. Further, the lack of standard data formats and protocols remains a significant barrier to the full integration of AI, LCC, and BIM [6].

Additional concerns arise at the regulatory and professional levels. The inherent opacity of AI models (commonly referred to as the black-box issue) and the lack of standardized validation frameworks reduce trust and slow down adoption among AEC practitioners [14].

Despite these limitations, the integration of BIM, AI, and LCC holds considerable potential to automate and enhance cost management, promoting greater consistency, transparency, and traceability across the project life cycle. While previous research has primarily focused on either the BIM–LCC or the AI–LCC stream, there is an increasing imperative to explore how these two research trajectories can intersect and evolve into a unified, intelligent cost-management framework. By combining information modeling, predictive analytics, and life-cycle costing, the AEC sector can move toward more resilient, efficient, and sustainable project-delivery strategies [6,12].

This article aims to contribute by systematically mapping and critically comparing two parallel research streams, LCC–BIM and LCC–AI, and by identifying critical obstacles and future synergies that could enable the more comprehensive integration of all three components. The goal is to map emerging synergies, identify methodological gaps, and define future research directions to advance digital enabled, intelligent, and integrated cost-management practices in the construction industry.

This study seeks to address the following research questions:

• RQ1: What are the main methodologies developed for integrating LCC and BIM, and what operational limitations hinder their large-scale adoption in the AEC sector?
• RQ2: How can AI techniques support the improvement of LCC practices, and what barriers limit their practical application?
• RQ3: What are the key obstacles and future synergies that could enable the combined integration of BIM, LCC, and AI, and how can this convergence foster the development of data-driven cost-management strategies?

2. Tools and Methodology

In order to systematically consolidate the current state of the art, identify existing research gaps, and highlight future research opportunities, the adoption of a Systematic Literature Review (SLR) represents a well-established methodological practice [15], following the PRISMA guidelines. The SLR follows a transparent, structured, and replicable process (Figure 1), with clearly defined criteria for selecting and analyzing sources. This methodology ensures the reproducibility of the review process by other scholars, thereby enhancing the robustness and reliability of the resulting findings.

The SLR process comprises a series of methodically structured and iterative stages, which are progressively refined until the desired level of comprehensiveness and reliability is achieved. For the retrieval of relevant literature, the Scopus database was utilized, as it is among the most comprehensive repositories encompassing a broad spectrum of peer-reviewed scientific publications. To complement the qualitative synthesis of the literature, a bibliometric analysis was also conducted using RStudio (R-4.2.2). This combined methodological approach enables quantitative visualization and mapping of the intellectual structure of the research field, facilitating the identification of key themes, influential authors, and emerging trends.

Figure 1 presents the methodological framework adopted in this study. The process begins with the formulation of the research questions, followed by the identification of relevant keywords for both research streams, as detailed in Section 2.1. Subsequently, a bibliometric analysis was conducted to explore quantitative metrics and emerging trends within the literature, as discussed in Section 2.2.

A structured literature review was then performed, using a stepwise selection approach to retain only the most relevant and high-quality contributions, as outlined in Section 2.3 and Section 2.4. The insights derived from both research streams were then integrated and examined to identify common themes, research gaps, and potential avenues for future investigation, as elaborated on in Section 3. This comparative synthesis enabled the definition of overarching research objectives that bridge both streams.

Drawing on these integrated findings, the study proposes specific research goals and outlines strategic directions for future research to inform and guide subsequent scholarly work.

2.1. Keywords Selection

At this stage, it is essential to identify the appropriate keywords to structure the search query. Initially, all the keywords (Figure 2) were combined into a single, comprehensive query. However, this approach failed to yield any relevant articles from the databases. As a result, it was considered methodologically appropriate to conduct two separate literature searches, each representing a distinct research stream. This strategy enabled a subsequent comparative analysis to determine whether any overlapping results emerged that could suggest points of convergence and support the potential integration of the two lines of inquiry. Both streams share LCC as a central theme, with the analyses respectively focusing on its integration with AI and BIM.

2.2. Bibliometric Analysis

The bibliometric analysis was conducted to extract quantitative data from the selected body of literature.

As illustrated in Figure 3, which compares the annual scientific output based on LCC–BIM and LCC–AI topics, both research streams have shown an overall growth trend over the past decade, reflecting the increasing academic interest in the associated topics, such as life-cycle cost, AI, and BIM. The number of publications related to LCC–BIM began to rise significantly in 2017, reaching a peak between 2019 and 2023, a period marked by the wider adoption of BIM in the AEC sector. In contrast, LCC–AI studies grew more gradually, followed by a pronounced acceleration beginning in 2020, likely influenced by the expansion of AI applications in the construction industry.

While the LCC–BIM topic appears to be more established, LCC–AI is emerging as a dynamic and fast-growing research area. Notably, in 2024, the number of publications focusing on LCC–AI surpassed those centered on LCC–BIM for the first time, signaling a shift in scholarly interest toward data-driven cost-estimation approaches. The apparent decline in 2025 is not indicative of a real reduction in scientific production. Rather, it reflects the temporal limitation of the bibliometric analysis, which was conducted in February 2025 and thus captures only the early months of the year.

Indeed, early data from 2025 already suggest continued growth: by February, 14 publications on LCC–BIM and 22 on LCC–AI had already been indexed. To provide a more accurate estimate of the year’s total output, a linear regression model was applied to past annual publication data (2022–2024), using Equation (1):

Publications _year = (a · x) + b

(1)

where “x” is the year index (2022 = 0, 2023 = 1, 2024 = 2, 2025 = 3), “a” is the slope (yearly growth rate), and “b” is the intercept (baseline number of publications in 2022). The slope “a” and intercept “b” are calculated using least squares fitting based on the observed values for 2022–2024. This model projects approximately 45 LCC–BIM and 57 LCC–AI publications for 2025, reflecting both the baseline trend and the acceleration observed in AI-driven research over the past two years.

These projections suggest that both research areas are poised to sustain or even surpass previous publication levels by year’s end, thereby reinforcing the ongoing consolidation and expansion of digital methodologies for LCC analysis in the AEC industry.

Figure 4 and Figure 5 show the geographical distribution of scientific publications by country across the two research streams. The maps reveal a distinct global pattern, with both topics attracting contributions from a wide range of countries, but with some differences. As shown in Figure 4, research on LCC–AI exhibits a broader global dispersion, with notable activity in Asia, the Middle East, and Africa. Countries such as India, China, Iran, and Egypt are actively engaged in publishing on this topic. This suggests that AI is being explored as a versatile and accessible tool to improve cost analysis, even in regions where digital construction platforms like BIM are not yet fully implemented.

Conversely, Figure 5 shows that LCC–BIM research is more heavily concentrated in high-income countries with established digital ecosystems. The United States, the United Kingdom, Australia, and several Western European countries report the highest number of publications. This is likely due to the more widespread adoption of BIM standards and technologies in these regions, often driven by public policy and procurement requirements.

Some countries, such as China, India, and South Korea, feature prominently in both maps, confirming their strategic role in construction innovation and their dual investment in BIM and AI-based solutions for life-cycle cost management.

Figure 6 and Figure 7 display the most frequent keywords found in the LCC–AI and LCC–BIM literature, respectively. These keyword visualizations facilitate the identification of prominent research themes and emerging directions within each domain.

In Figure 6, which corresponds to the LCC–AI domain, the most common keywords are Artificial Intelligence, Life Cycle, Cost, Decision Support Systems, and Decision Making. These reflect the focus on AI-based tools that assist with early-stage evaluations, uncertainty management, and complex decision processes. Other frequent keywords like Machine Learning, Forecasting, Optimization, and Deep Learning indicate an increasing interest in data-driven models for predicting cost-related parameters.

In Figure 7, which illustrates the LCC–BIM stream, keywords such as Architectural Design, Life Cycle, Cost, Building Information Modelling, and Sustainable Development are the most prominent. These terms highlight a more design-integrated approach, where BIM is used not only for cost estimation but also to support sustainability and performance management across the project life cycle. Additional keywords, such as Environmental Impact, Life-Cycle Analysis, and Construction Industry, suggest broader applications of BIM in managing building complexity and long-term performance.

Taken together, these keyword distributions show how the two streams complement each other: while LCC–BIM research integrates cost and sustainability into digital design workflows, LCC–AI focuses on improving cost estimation and decision-making through intelligent systems.

Over time, both streams have seen a steady rise in keyword frequency, indicating growing maturity and interest in these fields. This trend confirms the increasing relevance of combining BIM and AI with life-cycle thinking to support sustainable, informed decision-making in construction.

2.3. Paper Filtering

At this stage, the resulting set of articles underwent a series of filtering criteria designed to minimize the inclusion of irrelevant studies in the subsequent phases of analysis. The following filters were applied:

• Inclusion was limited to peer-reviewed journal articles, to ensure the consideration of high-impact scientific contributions.
• Only articles written in English were retained, to enhance accessibility and comprehensibility for the broader scientific community.
• Selection was restricted to subject areas about Engineering, Environmental Sciences, and Computer Science, in alignment with the research scope.
• Only publications from 2015 onwards were included, to ensure a focus on recent and relevant developments in the field.

In the following phase, additional screening was carried out based on the relevance of article titles and abstracts, assessed in accordance with the predefined research questions.

2.4. Paper Selection and Analysis

At this stage, the full texts of the filtered papers were thoroughly reviewed to enable their categorization. Given the distinct nature of the two research streams, each was analyzed using tailored selection criteria. For the stream focused on the integration of LCC with BIM, only those articles presenting validated methodologies for incorporating BIM into LCC calculations were included. Conversely, for the stream concerning the integration of LCC with AI, only studies that explicitly applied AI techniques to the computation of LCC were considered.

Both the selection and analysis procedures were customized for each stream. In the case of LCC–BIM integration, the review focused on identifying methodologies, calculation formulas, limitations, and advantages. For the LCC–AI integration, the analysis concentrated on the adopted AI methodologies, as well as their respective limitations and benefits.

As an outcome of this process, a theoretical framework was developed to synthesize and integrate the findings from both streams of the literature review.

3. Results

3.1. LCC–AI

As evidenced by the limited number of studies selected and analyzed, the integration of LCC and AI continues to represent a largely unexplored area within the academic literature. This notable lack of contributions highlights a significant research gap and underscores the need for further investigation to advance digital cost-management practices in the construction sector.

3.1.1. Methodological Approaches to the Integration of LCC and AI

The reviewed studies demonstrate the application of diverse AI-based methodologies to support sustainability assessments and cost evaluations in engineering and product development. The four primary studies reviewed include Teodosio et al. [16], Malviya et al. [17], Singh and Sarkar [18], Cerchione et al. [13].

Teodosio et al. [16] presents the MOUNDS algorithm, which is a deep-learning-based tool developed to simultaneously predict embodied energy, greenhouse gas emissions, life-cycle cost, and deflection of waffle and stiffened raft slabs compliant with Australian Standard. It leverages life-cycle inventory data, ISO-compliant cost analysis, and hydro-mechanical finite element simulations, using datasets split into training, validation, and testing sets. By modeling the complex interplay among environmental, economic, and structural parameters, MOUNDS supports informed decision-making in the early stages of slab design.

Yuan et al. [6] investigate the application of Natural Language Processing (NLP) techniques to extract economic data from technical specifications and contracts.

Malviya et al. [17] apply a multi-objective Particle Swarm Optimization (MOPSO) approach to minimize the LCC and environmental impact of NiZn batteries by selecting optimal combinations of countries for raw material sourcing and waste disposal. An AI-based framework, integrating data from the Ecoinvent database, OpenLCA software (v1.11.0), and other public sources, supports the optimization process. The results were qualitatively validated using the Analytic Hierarchy Process (AHP) and expert input, revealing a strong alignment between AI-driven and expert-based country selections.

Singh and Sarkar [18] introduce the development of an AI-based tool utilizing an Artificial Neural Network (ANN) to predict the LCC and Life-Cycle Assessment (LCA) of products during the early design phase. The tool features a user-friendly graphical interface for inputting key design parameters—such as product size, material density, manufacturing process, transport mode, and recyclability—while the back-end ANN, trained on 200 life cycle inventory cases, calculates the corresponding carbon footprint and cost outputs with up to 95% accuracy. Designed to support decision-making in environmentally conscious product development, the tool significantly reduces the time and data burden compared to conventional LCA/LCC approaches.

Collectively, these methodologies underscore the increasing integration of AI and Machine Learning (ML) with life-cycle thinking to support early-stage design decisions and streamline complex sustainability assessments. Typically, these approaches follow a common sequence for LCC estimation: they begin with the collection of a large, product-specific dataset (with properties to calculate LCC), which is then used to train an AI or ML algorithm capable of automatically predicting the LCC of new products based on a defined set of input parameters.

3.1.2. Benefits and Limitations

AI- and ML-based methods offer several notable advantages for LCC estimation. Chief among them is the ability to compute LCC values automatically, even in the absence of a complete dataset, thus simplifying the assessment process. Given the high dimensionality and complexity inherent to sustainability evaluations, these methods are particularly well-suited to optimization tasks, where AI consistently outperforms traditional analytical techniques [13,16,17]. However, a major limitation lies in their dependence on large, high-quality training datasets [18]. In the construction sector, this is further complicated by the heterogeneity of products and processes, which significantly impairs the generalizability of such models. In this context, Cerchione et al. [13] conducted an SLR to explore the integration of life-cycle methodologies with quantitative techniques aimed at supporting sustainability assessments. They emphasize the lack of applications involving emerging technologies such as AI and blockchain and underscore the need for further methodological innovation. Another significant gap identified is the lack of integration with Building Information Modeling (BIM) across the reviewed studies [13]. Despite BIM’s potential as a centralized data environment for automating and enhancing life-cycle assessments, none of the AI- or ML-based approaches incorporate BIM into their frameworks.

This oversight underscores a critical research opportunity: bridging digital design environments and AI-driven life-cycle analysis to enable more streamlined, data-driven, and integrated decision-making processes within construction workflows.

3.2. LCC–BIM

The analysis of the integration between LCC and BIM has stimulated the development of a variety of approaches aimed at overcoming the limitations of traditional LCC tools. These tools are often hindered by operational complexity, limited interoperability, and fragmented data, which impair consistent and efficient cost management. In this context, the growing adoption of BIM offers a strategic opportunity to address these challenges by centralizing multidisciplinary information within a shared digital environment.

3.2.1. Methodological Approaches for the Integration of LCC and BIM

The literature review identifies two main approaches for LCC–BIM integration: (1) embedding or linking LCC data within the BIM model, and (2) extracting quantities (Quantity Take-Off—QTO) from the BIM model for external analysis.

The first approach (1) aims to transform the BIM model into a rich information container capable of managing cost data throughout the asset’s life cycle. By embedding or linking LCC parameters directly in the BIM environment, this strategy improves the traceability, updateability, and coordination of economic information across disciplines and project phases.

The literature identifies two main operative modes within this approach:

• Embedding cost data directly into BIM objects using custom parameters or classification systems.
• Linking external databases via APIs, plugins, or visual programming tools (e.g., Dynamo).

In the first case, cost-related information—such as acquisition, maintenance, service life, and disposal—is directly associated with BIM elements, enabling integration between design modeling and life-cycle cost analysis. This supports automated or semi-automated simulations during both design and operation phases. A notable example is BIMEELCA [7,19], developed to support integrated environmental and economic assessments by connecting external LCA and LCC data to BIM components. The tool aims to enhance interoperability between modeling and evaluation environments, reducing fragmentation between design and analysis.

This integration also facilitates continuous data updates by various stakeholders (e.g., facility managers), promoting a dynamic and centralized information system for sustainable and informed decision-making.

A variant of the previously described methodology involves the use of external databases connected to BIM authoring tools through APIs, plugins, and visual programming environments (e.g., Dynamo) [20,21,22]. In this configuration, economic data—such as unit costs, service lives, and maintenance rates—is stored externally and linked to BIM elements, enabling cost analyses based on parameters like the type and quantity without embedding information directly in the model [23]. For example, the Relational Database Management System (RDBMS) by Le et al. [24] integrates Revit with Microsoft Access, Excel, and Dynamo, automating data collection and processing through SQL queries. Similarly, the Integrated Building Life-Cycle Assessment Tool (IBLAT) links Revit to a SQL database containing LCC data specific to the Chinese context, enabling automatic cost assignment through dedicated plugins [8]. Almomani and Almutairi [2] propose a framework based on parametric BIM models for automating LCC analyses, while Marzouk et al. [4] emphasize how dynamically embedded cost data can facilitate scenario simulations and risk management.

Furthermore, recent research suggests the use of semantic modeling strategies to integrate LCC parameters into Industry Foundation Classes (IFCs), although further development is needed to achieve scalable, reusable, and interoperable solutions [7]. Moreover, Lu et al. [11] identify data quality and consistency as critical success factors for BIM–LCC integration. Yuan et al. [6] further stress the need for standardized associations between building components and economic parameters within openBIM environments, to ensure interoperability across project stages. These solutions highlight the potential of external databases to improve the scalability and updateability of LCC analyses in BIM workflows.

The second approach (2), which is more common, involves extracting quantities (QTO) from the BIM model and processing them in external environments. Typically, this approach is divided into two variants: the use of spreadsheets (e.g., Excel) or dedicated LCC software.

Excel-based methods offer simplicity, transparency, and flexibility; however, they have limitations, especially in terms of automation and data management. The lack of automatic integration between the BIM model and the spreadsheet requires repetitive manual operations, making it prone to error and time-consuming. This mode is common in small to mid-scale projects, where the simplicity of the tools prevails over more complex automation requirements. For instance, Rostamiasl and Jrade [25] propose a BIM–Excel framework to assess LCC in aging-in-place housing. However, this method is prone to manual errors and lacks automation, requiring repeated updates when design changes occur.

A more advanced alternative involves using specialized external platforms to perform LCC analysis [1,8,10,26,27]. Tools such as Legap [28], IES VE/IMPACT [29], ATHENA Impact Estimator, SimaPro, and CostLab process quantitative data exported from the BIM model [7,19,30]. These tools support more complex simulations by integrating economic and environmental parameters and employing sophisticated models (e.g., Monte Carlo simulations, genetic algorithms), but their integration with BIM environments is limited. Workflow fragmentation, the need for multiple licenses, and the requirement for technical expertise represent major constraints. Solutions such as RenoDSS and data exchanges with external cost databases (e.g., Molio) through Dynamo VPL scripts illustrate efforts to partially bridge these environments [31].

The effectiveness of these approaches depends heavily on the quality of the data flow between tools, which must be continuous and accurate to prevent errors in calculations and simulations. Managing information flow becomes critical in complex scenarios involving numerous LCC parameters, where consistency and timely updates are essential.

In some cases, BIM models are extended beyond cost estimation to support energy simulations using Building Energy Models (BEMs). This integration allows for the inclusion of operational energy costs within LCC analyses, improving the realism of long-term economic forecasting. Studies by Alizadeh et al. [32], Pučko et al. [27], and Jiang et al. [33] demonstrate the benefits of integrating BIM–BEM workflows.

Further contributions, such as those from Istanbul Technical University [34], have explored the energy analysis of reference buildings to accurately determine thermal loads for heating and cooling. Similarly, Jausovec and Sitar [28] or Tushar et al. [35], use regression models and sensitivity analyses to optimize both energy and cost efficiency.

These practices confirm the growing trend towards a holistic approach to LCC analysis, where energy performance is no longer treated as an independent variable but as an integral part of the economic evaluation process.

3.2.2. Benefits and Limitations

The integration of LCC and BIM offers a wide range of advantages for the AEC sector, both at operational and strategic levels. One of the most significant benefits lies in the centralization and interoperability of data. BIM allows for the consolidation of geometric, material, performance, and economic information within a unified model. This integrated structure enables coordinated workflows among disciplines and reduces data redundancy, improving overall project consistency and collaboration [29]. The result is a more fluid and collaborative process, with a shared and constantly updated information base.

Moreover, automating QTO directly from the BIM model enhances both the speed and accuracy of cost estimations. By minimizing manual input, this process reduces the likelihood of human error and streamlines early-stage budgeting [32]. A further advantage of this integration is the dynamic nature of the BIM environment, where design modifications are automatically reflected in the associated cost data. This facilitates real-time updates of life-cycle cost calculations and supports the comparison of alternative design scenarios [8].

Another key feature is the ability to visualize economic data directly within the BIM model. Through color-coded maps or graphical overlays, stakeholders can intuitively understand the cost implications of specific design choices or materials. This improves communication, particularly with non-technical participants, and enables more inclusive and transparent decision-making processes [23].

Thanks to its parametric modeling capabilities, BIM also supports life-cycle-oriented evaluations by enabling early and accurate comparisons of different design alternatives based on their total cost, including construction, operation, maintenance, and disposal [29]. The enrichment of BIM models with LCC information fosters stronger stakeholder engagement by providing a shared, data-driven language across project teams. This facilitates alignment between technical and financial objectives and supports more informed and coordinated decisions.

Despite these advantages, several challenges still hinder the widespread and effective implementation of LCC–BIM integration. One of the primary issues is the methodological heterogeneity observed in the literature, which reveals persistent difficulties in standardizing workflows and ensuring interoperability across diverse software platforms. The use of fragmented toolchains, where modeling, cost estimation, and sustainability analysis are carried out in separate environments, limits automation, increases operational complexity, and raises the risk of inconsistent data flows. Indeed, the integration of BIM and LCC data within the project database is considered one of the main challenges [36].

A particularly critical limitation concerns the integration of economic data within open standards such as IFC [37] or Construction Operations Building Information Exchange (COBie) [38]. Although these standards offer promising pathways for data exchange across proprietary platforms, the structuring of cost-related attributes remains underdeveloped. IFC lacks dedicated and widely implemented classes and property sets for core LCC parameters, such as acquisition, maintenance, and replacement costs, with IfcCostItem being the only relevant class. However, it is rarely applied in practice. Similarly, COBie is not designed to represent structured cost information. As a result, structured mappings between BIM objects and cost data are seldom adopted, and currently, there are no commercial tools that support this integration in a standardized manner. Most workflows rely on custom extensions or external linking strategies, which compromise interoperability, limit scalability, and reduce repeatability. These limitations often result in data loss or semantic inconsistencies during software transfers, thereby undermining the reliability of BIM-based LCC assessments [7]. To overcome this gap, there is a pressing need for the development of standardized cost ontologies and interoperable workflows that can enable the seamless and verifiable integration of economic data into openBIM environments.

Another major challenge involves addressing uncertainty. Many current BIM-based LCC approaches rely on deterministic assumptions and fail to account for the stochastic nature of key variables, such as cost variability [8,10], inflation, and discount rates [4,9,10]. This lack of probabilistic modeling limits the robustness and credibility of life-cycle cost assessments, especially in early design phases where uncertainty is highest.

To ensure reliable and effective LCC–BIM integration, it is essential to maintain continuous, accurate, and transparent data flows across systems and stakeholders. Recent studies have begun to address these limitations by exploring the integration of advanced computational techniques—such as artificial intelligence, optimization algorithms, Monte Carlo simulations, and multi-criteria decision-making—within BIM environments [4,9,10,33,39]. Although these approaches are still in a developmental phase and not yet widely adopted in practice, they offer promising directions for enhancing the predictive and analytical capabilities of LCC assessments.

4. Discussion

The systematic review reveals an expanding but disjointed research landscape regarding the integration of LCC, BIM, and AI in the AEC sector. While the individual advantages of each technological domain are extensively substantiated, their synergistic deployment remains underexplored. This study systematically evaluates and juxtaposes two parallel research streams, LCC–BIM and LCC–AI, outlining the factors that hinder integration into an integrated framework. The findings confirm that BIM–LCC integration has achieved greater methodological maturity, whereas AI-based LCC approaches are still in their early stages and largely confined to experimental or product design contexts.

From a methodological perspective, LCC–BIM integration has evolved along two main trajectories: (1) embedding or linking cost databases within BIM environments, and (2) extracting data from BIM models for external LCC analyses. Both approaches offer significant potential for enhancing data interoperability, transparency, and efficiency in cost assessments. However, important challenges persist in terms of data standardization, information flow consistency, and workflow adaptation across different project requirements and software ecosystems. Despite the evident operational advantages, including automated data extraction, real-time updates, and the visual representation of economic impacts, practical implementation is often constrained by the lack of interoperable tools, robust cost databases, and shared modeling protocols.

In contrast, the application of AI to LCC remains largely experimental and unstructured. The reviewed studies demonstrate the feasibility of applying Machine Learning algorithms, such as Neural Networks and optimization techniques, to predict LCC and support sustainability-driven decision-making. However, these methods rely heavily on large, high-quality datasets and are rarely applied to building-specific or construction site contexts. Notably, no existing AI-driven approach reviewed in this study incorporate BIM, confirming that the structured integration of LCC, BIM, and AI has not yet been realized. This absence represents a critical gap with both technical and cultural origins. From a technical standpoint, AI models require standardized, structured, and labeled datasets that BIM environments could potentially provide, but which are rarely made available in accessible formats. From a cultural and professional perspective, the fragmentation between the actors involved hinders interoperability and collaboration between different domains. Moreover, concerns about data ownership, model explainability, and software compatibility further discourage experimentation in operational settings. These issues significantly limit the potential of AI to enhance and automate LCC analyses within BIM-enabled environments.

The comparative analysis of the two research streams reveals that, although they follow separate development paths, several methodological and technological synergies remain underexploited. For example, BIM could serve as a structured data environment for feeding AI models with consistent, standardized, and project-specific inputs. Conversely, AI could support BIM-based workflows by improving the reliability and robustness of LCC parameter estimation. Unlocking these synergies requires targeted research efforts and a shift from siloed experimentation to more integrated approaches.

It is essential to clarify that the goal of integrating AI into LCC workflows is not to redefine the underlying cost models, which are well-established in the literature, but rather to enhance the estimation of input parameters that significantly influence LCC outcomes. Sensitivity analyses have consistently demonstrated how uncertainties in variables such as the service life, maintenance frequency, and replacement costs propagate through LCC calculations and significantly affect final outputs [4,9,34,35].

These uncertainties often originate from upstream sources, including volatile market conditions, regulatory shifts, and environmental variability [8]. In this context, AI techniques are particularly valuable for identifying hidden patterns and correlations within historical data that are otherwise difficult to capture using conventional models. By learning from past behaviors and interactions, AI can contribute to more robust, data-driven, and probabilistic input estimates.

Among various AI methods, Bayesian Networks (BNs) emerge as especially promising for LCC applications. Their suitability stems from several distinctive features: the ability to represent uncertainty probabilistically, the capacity to operate effectively even with incomplete or noisy data, and the possibility to model causal relationships between interdependent variables that are not governed by deterministic rules. These characteristics make BNs highly appropriate for modeling the dynamic, uncertain, and multi-variable nature of construction projects throughout their life cycle [40,41]. Compared to black-box methods like deep learning, BNs also offer greater interpretability, which is crucial in cost-sensitive environments where both practitioners and public authorities require transparency and explainability. This aspect is particularly relevant in LCC applications, where stakeholders, including public bodies and investors, demand traceable and auditable models. While high-performing black-box models often lack interpretability and hinder validation, Bayesian approaches enable a clearer understanding of how predictions are generated, thereby supporting greater trust and accountability.

Another critical concern is the interpretability and validation of AI models. As emphasized in the literature, the opacity of many predictive algorithms and the lack of shared regulatory frameworks pose major obstacles to adoption in professional practice. The so-called ‘black-box issue’ becomes especially problematic in high-stakes projects, where cost decisions must be justified with transparent, verifiable evidence. These issues are particularly acute in high-stakes or public projects, where traceability, auditability, and stakeholder confidence are essential.

In conclusion, this review highlights not only the current progress and limitations of LCC–BIM and LCC–AI integrations, but also the untapped potential of their convergence. The combined integration of BIM, AI, and LCC represents a promising yet complex research frontier. To translate these findings into practical advancements, it is essential to consider how such insights can be operationalized across real-world scenarios.

4.1. Implications for Practice

The findings of this review provide actionable insights for enhancing digital cost management in the AEC sector. First, the progressive integration of BIM, LCC, and AI offers a concrete pathway to increase cost transparency and long-term decision reliability. In particular, BIM models can serve as a structured data source for training predictive AI systems such as Bayesian Networks, which can be used to infer uncertain or missing LCC parameters based on historical project data.

This convergence can support practical use cases, such as the following:

• Public procurement workflows, where life-cycle costs must be forecasted and justified ex ante and where traceability and explainability are required by law.
• Sustainable building design, where BIM–LCC–AI integration can simulate the long-term environmental and economic impacts of different material and design choices in the early stages.
• Predictive maintenance and facility management, where AI-enhanced LCC can be embedded in BIM-based digital twins to optimize life-cycle interventions and reduce the total cost of ownership.

Building on these practical implications, the following section outlines a structured research roadmap to guide the development of integrated and intelligent LCC–BIM–AI systems.

4.2. Future Works

To move toward a unified and intelligent cost-management framework, future research should be structured through three progressive stages:

(1)

Data Structuring and Standardization: This initial stage should focus on the development of open, machine-readable datasets, the definition of standardized ontologies for economic attributes, and the semantic enrichment of BIM environments to support structured cost information. This includes improving the representation of LCC parameters within IFC models and aligning them with AI-ready data schemas.

(2)

Integration and Technical Interoperability: Once data foundations are established, the next step involves technically integrating BIM and AI workflows. This requires developing APIs, data pipelines, and middleware to allow seamless interaction between BIM platforms, cost estimation tools, and AI engines. AI can be progressively introduced to automate parameter estimation, scenario analysis, and anomaly detection.

(3)

Application and Validation in the Real-World: In the final phase, integrated BIM–LCC–AI frameworks should be tested in pilot projects across different building types and procurement settings. Emphasis should be placed on explainability, auditability, and usability to ensure adoption by practitioners. Benchmark datasets and standard validation protocols (e.g., uncertainty metrics, calibration procedures) should be developed to compare models and foster trust.

These stages are not strictly sequential but can be pursued iteratively and in parallel across different research communities. Ultimately, this roadmap aims to guide the transition from siloed digital solutions toward cohesive, intelligent, and transparent systems for life-cycle-oriented cost management in the AEC sector.

At the strategic level, this integration requires resolving several interrelated challenges: developing standardized ontologies and schemas for economic data, improving access to granular, high-quality, and open datasets, and better aligning technological tools with established professional workflows. At the same time, institutional and educational initiatives will be needed to promote digital competencies, support data literacy, and define shared validation and certification protocols to foster trust in AI-assisted cost management systems.

5. Conclusions

This study has systematically reviewed the current state of research on the integration of LCC, BIM, and AI in the AEC sector. The main contribution of the study lies in mapping two distinct but converging research trajectories (LCC–BIM and LCC–AI) and in identifying both the technical and methodological challenges that prevent their convergence. The results demonstrate that LCC–BIM approaches have achieved a higher level of methodological maturity, with growing applications in project planning and evaluation. In contrast, LCC–AI remains largely exploratory, with few implementations in the construction context.

Regarding RQ1, the analysis has demonstrated that BIM-enabled LCC estimation benefits from improved data structuring, automated quantity takeoff, and dynamic cost updates. However, it also suffers from poor interoperability across platforms and a lack of standardized protocols for integrating economic data into openBIM schemas, such as IFC.

In response to RQ2, the potential of AI in LCC processes is evident in its ability to address uncertainty and enhance parameter estimation through predictive models. Bayesian Networks offer interpretability and probabilistic reasoning. Nevertheless, the reviewed studies lack application in real-world construction workflows and often ignore BIM as a structured data source.

As for RQ3, the review identifies a promising yet underdeveloped opportunity: the combined integration of BIM, LCC, and AI. This synergy could address current limitations in early-stage decision-making, scenario simulation, and life-cycle risk assessment. Integration is especially relevant in high-stakes domains such as public procurement, green building design, and predictive maintenance.

To unlock this potential, the study proposes a roadmap articulated across three progressive stages: (1) data structuring and standardization of LCC information in BIM; (2) technical integration of BIM–AI workflows via APIs and interoperable formats; and (3) real-world validation through pilot projects, including the development of open datasets and benchmark metrics.

From a practical standpoint, these findings offer actionable recommendations for both researchers and industry stakeholders. BIM models should be leveraged as training environments for AI algorithms, enabling the development of adaptive cost-estimation tools that can be applied to various projects. Future systems should support real-time feedback, explainability, and scenario testing, integrated directly into BIM platforms via user-friendly interfaces. Institutions and public agencies play a crucial role in fostering open innovation by promoting the availability of structured data and shared evaluation protocols.

In conclusion, while the current literature remains fragmented and often conceptual, the integration of BIM, AI, and LCC holds transformative potential. Advancing from isolated applications to intelligent, interoperable, and transparent cost-management frameworks requires not only technological innovation but also cross-disciplinary collaboration and institutional commitment. The path is open for a new generation of research that is not only rigorous but also operationally impactful for the digital transformation of the AEC sector.