Integrating IoT and AI for Sustainable Energy-Efficient Smart Building: Potential, Barriers and Strategic Pathways

Abstract

The global drive toward sustainability and energy efficiency has accelerated the development of smart buildings integrating the Internet of Things (IoT) and Artificial Intelligence (AI). These technologies optimise energy use, enhance occupant comfort, and advance building management systems. This study examines the integration of IoT and AI in energy-efficient smart buildings, emphasising applications and challenges. A qualitative methodology, combining systematic literature review, case study analysis, and systems analysis, underpins the research. Findings indicate that IoT enables smart metering, real-time energy monitoring, automated lighting and HVAC, occupancy-based energy optimisation, and renewable energy integration. AI complements these functions through predictive maintenance, energy forecasting, demand-side management, intelligent climate control, indoor air quality automation, and behaviour-driven analytics. Together, they reduce carbon emissions, lower operational costs, and improve occupant well-being. However, challenges remain, including data security and privacy risks, interoperability gaps, scalability and cost constraints, and retrofitting difficulties. To address these, the paper proposes a systems thinking-enabled conceptual framework structured around three pillars: adopting IoT and AI as enabling technologies, overcoming integration barriers, and identifying application areas that advance sustainability in smart buildings. This framework supports strategic decision-making toward net-zero and resilient building design.

1. Introduction

1.1. Sustainability in the Built Environment and Buildings

Sustainability in the built environment encompasses the environmental, social, and economic aspects of urban spaces that are energy-efficient, resilient, inclusive, and in a position to provide a high level of human comfort. The rapid process of urbanisation has made buildings and infrastructure the focus of sustainability debates due to their massive demand for resources and high carbon emissions. Buildings are responsible for approximately 45% of total energy use and can account for as much as 80% of global water consumption [1], making them one of the primary causes of environmental degradation [2]. Sustainable construction has emerged as a response to these concerns, featuring energy-efficient systems, water conservation, and novel materials such as high-performance concretes and insulation from bio-based products [3,4].

Sustainable buildings have, therefore, emerged as the cornerstone of such a transition. With a design view to minimising environmental impacts and improving the quality of life, they reduce emissions, optimise resource use efficiency, and manage the wastes effectively during the building’s lifetime.

Although costing more upfront, sustainable buildings save money over the long term [5], while improving health and comfort, as demonstrated by Milan’s Bosco Verticale and Amsterdam’s The Edge [3]. Emerging technologies, including energy management systems and Internet of Things (IoT), enable predictive maintenance and further extend these advantages to enhance building durability and efficiency [4]. Despite these advantages, substantial barriers remain. High initial investment requirements, complex regulatory frameworks, and a lack of public awareness hinder the widespread adoption of this building category [6]. To overcome these limitations and continue advancing sustainable construction practices, several complementary strategies have been proposed.

Passive design, renewable energy integration, and the use of recycled or renewable materials are some strategies that further advance the performance of buildings. At the same time, it is increasingly recognised that beyond ecological efficiency, sustainability encompasses social equity and economic opportunity as well. Equitable urban design cultivates inclusivity and links sustainability to social and economic well-being [7,8]. Planning approaches, such as mixed-use zoning, help foster diverse, connected, and resilient communities [9].

In this context, decision-making frameworks play a critical role in advancing these goals. Tools such as geographic information systems (GIS) and urban economic models enable the integration of environmental, social, and economic data into long-term planning strategies [9]. A systems-thinking approach, which accounts for interactions among infrastructure, social dynamics, and economic forces, is now widely recognised as essential for building resilience in the face of climate change and global uncertainties [10].

Furthermore, several persistent challenges, including high capital needs, stakeholder complexity, and undervalued lifecycle benefits [8,11], necessitate holistic, integrated, and interdisciplinary approaches to embed sustainability throughout the building life cycle [12]. In this regard, the integration of IoT and AI offers new opportunities to optimise energy use, reduce emissions, and improve occupant well-being, enabling smart and sustainable urban futures.

1.2. Smart Sustainable Buildings and Energy Efficiency

Smart sustainable buildings extend sustainability principles by embedding digital technologies that optimise energy use, comfort, and overall performance [13]. By incorporating digital intelligence, smart buildings enable automation, real-time monitoring, and predictive control of resources.

Central to this transformation is the convergence of the IoT and AI or AIoT, where data from connected devices is analysed to inform intelligent decision-making [14]. Building Management Systems (BMS) integrate heating, ventilation, and air conditioning (HVAC), lighting, and security [15], while emerging self-supervised learning technologies enable continuous optimisation for energy and comfort [16]. These innovations reduce energy demand and operational costs, while also enhancing safety, health, and sustainability [17,18]. At a broader scale, smart sustainable buildings function as nodes within smart cities, contributing to efficient resource distribution and urban resilience [13,19]. Advances in adaptive design, innovative materials, and certification standards are further aligning them with global sustainability goals [17,20].

Energy efficiency is central to smart, sustainable building design [21,22]. IoT-enabled sensors, intelligent HVAC systems, and renewable integration ensure energy is used only when needed [23,24,25,26].

The adoption of renewable systems such as solar photovoltaics further reduces dependence on fossil fuels [21,27]. Energy harvesting technologies that capture ambient light, heat, or vibrations also contribute to energy autonomy [21,25]. These developments deliver both climate benefits and economic value, reducing emissions and operating costs [23,28]. However, barriers—including high capital costs, technological complexity, and limited awareness—continue to hinder adoption [26,28].

1.3. IoT and AI as Transformative Technologies

IoT and AI together create adaptive systems that optimise resources and minimise environmental impact through real-time monitoring, predictive maintenance, and intelligent energy management [29,30,31,32]. IoT devices such as smart meters and sensors capture operational data, while AI algorithms analyse it to automate control, forecast maintenance, and optimise consumption [29,30,31,32].

Digital twins that integrate IoT and AI enable the simulation and optimisation of building performance, supporting Net Zero Energy Buildings (NZEB) [33,34,35,36,37,38]. Despite their promise, adoption remains limited by high costs, technical and privacy challenges, and regulatory gaps [36,39,40]. Addressing these barriers requires robust policies, effective cybersecurity protocols, and enhanced stakeholder collaboration to ensure the secure and equitable deployment [39].

1.4. A Systems Thinking Perspective

The integration of IoT and AI in smart buildings reflects a systems-thinking approach that recognises interdependencies and feedback loops within complex building ecosystems [40]. Rather than treating building elements in isolation, systems thinking views them as interacting subsystems evolving over time. As illustrated in Figure 1, IoT devices generate data that AI analyses to support decision-making [14]. AIoT then informs building design, influenced by policy incentives, which shape energy management strategies. Effective energy management enhances efficiency, a defining feature of smart buildings, while barriers such as cost or skills shortages can hinder progress. Ultimately, smart buildings contribute to urban resilience, feeding lessons back into design and management for continuous improvement [36].

1.5. Research Gap, Objectives, and Questions

Despite rapid advances in the digitalisation of the built environment, there are significant research gaps regarding how the integration of IoT and AI can be systematically leveraged to enhance the energy efficiency and sustainability of smart buildings. Most existing studies have focused either on isolated IoT-enabled applications, such as monitoring, smart metering, and predictive maintenance, or AI-based optimisation and fault detection without considering their combined and synergistic potential, as captured by the concept of the Artificial Intelligence of Things (AIoT) [19,41,42,43,44]. The literature remains fragmented with respect to conceptualising the feedback dynamics between technology, human behaviour, and policy, which limits a comprehensive understanding of how these systems interact in delivering sustainable outcomes [43,44]. In particular, economic, technical, and institutional barriers, such as high upfront costs, interoperability challenges, and data privacy risks, are often considered in isolation. For example, economic, technical, and organisational challenges, such as high implementation costs, interoperability issues, data security risks, complexity of integration, and limited coordination amongst stakeholders, are amongst the commonly identified barriers [43,44]. Yet, systematic approaches to addressing these interrelated barriers across diverse building contexts remain limited. Additionally, environmental and ethical considerations, including the sustainability of digital infrastructures themselves, the energy required for data processing, and responsible AI governance, are also underexplored [45,46]. Proposed solutions, such as cybersecurity protocols, green technology innovation, and collaborative stakeholder strategies, thus remain piecemeal and inadequately aligned within an integrated strategic framework [28,43]. Similarly, emerging tools like digital twins [45] offer tremendous potential but still require better conceptual grounding and theoretical consolidation to enable scaling up and sustainability.

Taken together, the key gaps that emerge from the literature include a limited understanding of the combined capabilities of IoT and AI beyond isolated applications, fragmented and unintegrated strategies for adoption, and insufficient exploration of how technical, economic, and institutional barriers interact across smart building contexts, along with underexplored environmental and ethical implications. These complex, interconnected challenges require a holistic systems-thinking approach that integrates technological innovation with policy, economic, and human dimensions to bring forth adaptive, energy-efficient, and sustainable building ecosystems. Driven by the global urgency to decarbonise the building sector—which accounts for a substantial share of total energy consumption—and the demonstrated potential of IoT–AI integration to significantly reduce energy use while enhancing occupant comfort [1,43], this study aims to bridge the conceptual and practical divide by developing a systems-based framework to guide the sustainable and large-scale adoption of IoT and AI in smart building design and operation.

Consequently, the objectives of the study are:

• To examine the potential and functional capabilities of IoT and AI in enhancing energy efficiency and sustainability.
• To identify and analyse state-of-the-art applications of IoT and AI technologies in smart building environments.
• To investigate key barriers to adoption in both new and existing infrastructures and to develop a systems-thinking-based framework for sustainable, energy-efficient smart buildings.

In this context, the research questions (RQs) investigated are:

• RQ1: What IoT and AI technologies are available, and what roles do they play in energy management?
• RQ2: How does their integration enable state-of-the-art applications for energy management in smart building systems?
• RQ3: What barriers limit their adoption, and how can a systems-thinking framework support the development of sustainable, energy-efficient smart buildings?

2. Research Methods

This study adopts a qualitative research design that combines a systematic literature review with the analysis of selected case studies. A detailed research protocol guided both components, ensuring rigour, consistency, and minimisation of bias.

2.1. Research Protocol

The research protocol defined the study’s scope, research questions, search strategy, inclusion/exclusion criteria, data extraction, quality assessment, case study selection, and analytical methods, providing a framework to ensure methodological transparency and reliability (

Table 1. Research Protocol.

Item	Details
Research Questions	RQ1: What are the potential IoT and AI technologies and their role in energy management in smart buildings? RQ2: How does the synergetic integration of IoT and AI lead to state-of-the-art applications for energy management in smart building systems? RQ3: What are the major barriers to IoT and AI integration, and how can a systems-thinking framework foster the development of sustainable, energy-efficient smart buildings?
Database used	Scopus and Web of Science
Publication period	2001–2025
Keywords	IoT, AI, Sustainable, Energy-Efficient, Energy management, Smart Building, Green Building, Potential, Barriers, Strategies, systems thinking
Timeframe for literature search	May 2023–June 2025
Inclusion criteria	Population: Peer-Reviewed Journals, Books, Book Chapters, Theses, Conference Proceedings Interventions: IoT and AI technologies, Roles, Applications Context (Comparison): Buildings, Urban Areas, Cities, Global South, Global North Outcomes: Sustainability, smart, efficiency, and Resiliency. Others (Language): English, French, Spanish, or Portuguese
Exclusion criteria	Non-Peer-Reviewed Articles, Patents, Laws, Treaties Not aligned to energy management and smart buildings, hardware and software, Selective Reporting, Specific Contextual Studies, Qualitative Observational Studies, Blogs/Opinions Without Evidence
Data extraction	Used a standardised form (spreadsheet) to capture all relevant data
Quality assessment	Used the 27 PRISMA checklist to assess methodological quality, Risk of Bias (ROB2) analysis and Cohen’s Kappa analysis
Case studies	Four: The Edge in The Netherlands, Rinascimento III in Rome, Infosys’ corporate campus in India and Keppel Bay Tower, Singapore
Analytical approach	Used narrative and thematic analysis and synthesis of the data

).

2.2. Literature Review Process

The systematic literature review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework, ensuring rigour, reproducibility, and minimisation of bias [47,48,49]. It included peer-reviewed articles, books, conference proceedings, policy reports, and high-quality online sources on smart buildings, urban systems, energy efficiency, urban resilience, IoT and AI technologies, applications, barriers, strategies, systems thinking, and conceptual frameworks for energy management. The review was structured around PRISMA steps such as identification, screening, eligibility, and inclusion, and conducted between July 2024 and July 2025, with case studies identified concurrently from scholarly and publicly available sources. The PRISMA checklist is provided in Supplementary File S1 to ensure transparency, completeness, and reliability.

2.2.1. Search Strategy

The literature search drew on a diverse range of sources, including peer-reviewed journal articles, books, conference proceedings, policy reports, and credible web resources. Searches were conducted in Scopus and Web of Science, with search strings tailored to each database’s syntax and refined using predefined inclusion and exclusion criteria. In addition, relevant reports were identified through searches of the official websites of pertinent organisations. A representative search string is presented below:

“((IoT OR AI) AND (Sustainable OR Energy-Efficient) AND (Smart Building OR Green Building) AND (Potential OR Barriers OR Strategies)) AND (EXCLUDE (SUBJAREA, “PHAR”) OR EXCLUDE (SUBJAREA, “NEUR”) OR EXCLUDE (SUBJAREA, “ARTS”) OR EXCLUDE (SUBJAREA, “AGRI”) OR EXCLUDE (SUBJAREA, “CHEM”) OR EXCLUDE (SUBJAREA, “MEDI”) OR EXCLUDE (SUBJAREA, “BIOC”) OR EXCLUDE (SUBJAREA, “EART”)) AND (EXCLUDE (DOCTYPE, “cr”) OR EXCLUDE (DOCTYPE, “er”))”

A snapshot of the literature search using the search string in both databases is presented in Figure 2.

2.2.2. Screening Process

Screening was guided by the inclusion and exclusion criteria outlined in the research protocol (Table 1) and applied through the PICOSO framework (Population, Intervention, Comparator, Outcomes, Study Characteristics, and Others). The investigator and one research assistant independently reviewed titles and abstracts, followed by full-text assessment for eligibility. Any disagreements were resolved through team discussions, ensuring consistency and accuracy.

2.2.3. Data Extraction and Quality Appraisal

Data were extracted using a standardised form capturing publication details, objectives, methodologies, key findings, contributions, interventions, limitations, future research directions, and funding sources. The researcher and assistant independently extracted and cross-checked the data, which were compiled in Microsoft Excel for analysis. Quality assessment utilised the Cochrane Risk of Bias (RoB 2) tool and Cohen’s Kappa to evaluate methodological rigour and inter-rater reliability, with discrepancies resolved through discussion or arbitration by a senior reviewer. Most studies were assessed as low risk of bias, as detailed in the Supplementary Files (Study Characteristics (S2), RoB2 Results (S3), and RoB2 Summary (S4)), where File S2 provides key details of the included studies, File S3 highlight potential biases in individual studies, and File S4 synthesises the overall risk of bias to inform confidence in the review’s conclusions.

2.3. Case Studies Selection

To complement the literature review, four case studies were selected, including The Edge in the Netherlands, Rinascimento III in Rome, Infosys’ corporate campus in India, and Keppel Bay Tower in Singapore. These case studies were purposively selected to provide concrete illustrations of how IoT and AI technologies are being leveraged to enhance energy efficiency in smart buildings. The selection process drew upon peer-reviewed academic literature, reputable industry reports, and sustainability certification databases to ensure the inclusion of credible and representative examples. Cases were chosen to capture a broad spectrum of building types/locations, functions, IoT–AI integration features, energy savings/efficiency outcomes, cost and investment considerations, integration maturity levels, sustainability/certification outcomes, and implementation challenges, thereby enabling a comparative understanding of diverse technological and operational conditions.

The Edge in the Netherlands illustrates a pioneering smart office development that integrates advanced IoT and AI systems to optimise lighting, HVAC, and occupancy management, achieving a BREEAM Outstanding certification. Rinascimento III in Rome demonstrates the scalability of IoT-enabled monitoring and control in residential Net Zero Energy Buildings (NZEB), highlighting pathways for sustainable urban housing. Infosys’ corporate campus in India represents an enterprise-scale application of AI and IoT for predictive energy management, resulting in measurable reductions in operational energy demand across a large facility network. Keppel Bay Tower in Singapore illustrates the complexities of retrofitting existing high-rise commercial buildings with smart energy systems, culminating in the achievement of Super Low Energy (SLE) certification.

Collectively, these case studies span different building typologies, climatic regions, energy savings/efficiency outcomes, cost and investment considerations, and stages of technological maturity. They provide a holistic perspective on how IoT and AI integration can drive sustainability outcomes, offering insights into innovative practices such as digital twin deployment, AI-driven optimisation, and organisation-wide IoT management strategies.

2.4. Analysis and Synthesis

The analysis employed a deductive thematic approach, coding and grouping key concepts from the literature and case studies into themes aligned with the study’s objectives. Bibliographic analysis software, such as the Biblioshiny platform in R software (Version 2025.05.0+496) and VOSviewer (Version 1.6.20), were used to identify trends and perform thematic analysis. Themes and subthemes were coded based on the outputs from the software, supplemented by manual interpretation aligned with the focus and objectives of the study.

Themes included IoT and AI technologies for energy management, applications of IoT in building systems, AI applications in smart buildings, synergistic integration of IoT and AI (AIoT), barriers to adoption, and the role of systems thinking in addressing these barriers. Findings from the case studies were triangulated with insights from the literature to ground thematic patterns in practical contexts. This synthesis revealed both shared strategies and context-specific challenges, contributing to a more comprehensive understanding of how the integration of IoT and AI can support the development of sustainable, energy-efficient smart buildings.

It is to note that AI-assisted writing tools, specifically ChatGPT versions 4 and 5, were utilised to refine the language and enhance readability. Also, Scispace was used to access literature. All substantive content, including data interpretation, conclusions, and manuscript structure, was generated and validated by the author.

3. Results

The results of the analysis are organised into five dimensions: (i) characteristics of the reviewed documents; (ii) the potential of IoT and AI in smart buildings; (iii) synergistic integration of IoT and AI, with a focus on state-of-the-art AIoT applications; (iv) barriers to IoT and AI integration; and (v) discussion, strategic pathways, integrated framework, and policy implications.

3.1. Characteristics of Documents Used for Review

The database search identified 439 records from Scopus and 385 from Web of Science. After merging the datasets and removing duplicates using Biblioshiny in R software (Version 2025.05.0+496), 686 unique records remained for screening. Of these, 121 sources were selected for detailed analysis, including 66 journal articles (one of which was a preprint), 20 conference papers, and 26 books or book chapters. In addition, nine web articles or reports were included (

Table 2. Summary of Literature Reviewed.

Literature Source	Numbers	Share (%)
Journal articles	66	54.55
Conference Proceedings	20	16.53
Books/Book Chapters	26	21.49
Web articles/Reports	9	7.44
Total	121	100.00

). The overall selection process is illustrated in the PRISMA flow diagram (Figure 3).

The RoB2 assessment indicated a low risk of bias. The observed inter-rater agreement was 0.927, indicating that the raters agreed on 92.7% of the cases. After adjusting for agreement expected by chance, Cohen’s Kappa was 0.744 (ASE = 0.071), demonstrating a substantial level of agreement [50]. This agreement was statistically significant (T ≈ 9.075, p < 0.001), confirming that the concordance was unlikely due to chance (

Table 3. Level of Agreement between the Assessors (Risk of Bias).

Measure	Value	Std. Error	Approx. T	N of Valid Cases
Cohen’s Kappa (k)	0.744	0.071	9.075	111

). The final set of articles was subsequently used for the thematic analysis.

3.2. IoT and AI Technologies for Energy Management in Smart Buildings

Table 4. IoT and AI Technologies used in Energy Management in Smart Buildings.

IoTs	AI
Sensor networks, including temperature, occupancy, and energy usage sensors	Machine Learning (ML), Deep Learning (DL), and Reinforcement Learning (RL)
IoT-based Energy Management Systems (EMS).	Digital Twin, Edge Computing, Blockchain, and hybrid AI approaches.
Short-Term Load Forecasting (STLF) using K-Nearest Neighbour (KNN).	Automated Fault Detection and Diagnosis (AFDD) systems

presents the IoT and AI technologies applied in the energy management of smart buildings. Primary IoT technologies include sensor networks, such as those for temperature, occupancy, and energy usage [43,51], as well as IoT-based Energy Management Systems (EMS) [51,52]. These systems are often enhanced with forecasting techniques, for example, Short-Term Load Forecasting (STLF) using the K-Nearest Neighbour (KNN) algorithm [53].

The widely used AI technologies used in smart building energy management include Machine Learning (ML), Deep Learning (DL), Digital Twins, Blockchain, and hybrid AI approaches such as the firefly algorithm and genetic algorithms [37,38,54,55,56,57,58]. Reinforcement Learning (RL) is also applied, along with predictive maintenance models and Automated Fault Detection and Diagnosis (AFDD) systems [59].

3.3. The Potential of IoT and AI in Smart Buildings

3.3.1. IoT for Energy Management in Building Systems

The integration of IoT into building systems is transforming energy management by enhancing efficiency, reducing costs, and promoting sustainability [19,52]. Through real-time monitoring, automation, and intelligent controls, IoT reshapes how buildings consume and conserve energy [51,53,60,61]. Core applications include smart metering, automated lighting and HVAC, occupancy detection, and renewable energy integration, creating adaptive, energy-efficient environments aligned with global sustainability goals.

In this regard, Saleem et al. (2023) [62] demonstrate that integration of IoT and cloud computing enables real-time monitoring and effective energy management, particularly for air conditioning systems, which account for more than half of the total electricity consumed in Pakistan. The Smart Energy Management System (SEMS) implemented in buildings achieved energy savings of 15 to 49% by leveraging advanced algorithms and user-friendly interfaces to optimise energy usage and reduce energy costs. This real-world installation shows the scalability of the system for sustainable and energy-efficient smart grids.

Similarly, smart metering and real-time monitoring provide detailed insights into consumption patterns, allowing managers to identify inefficiencies and implement targeted energy-saving strategies [63,64]. When combined with predictive maintenance and demand–response mechanisms, IoT-enabled meters improve grid stability and reduce waste [65,66]. The automation of lighting and HVAC systems further enhances efficiency, as sensors dynamically adjust operations based on occupancy and ambient conditions, thereby reducing unnecessary energy use while maintaining comfort [43,63]. This approach lowers operational costs and minimises the need for manual intervention [67,68].

Occupancy detection is another critical component. Sensor-based systems align energy use with actual occupancy, optimising HVAC zoning, lighting schedules, and overall facility operations while enhancing security [49,51,65,69]. Moreover, IoT facilitates the integration of renewable energy and the coordination of smart grids, enabling the seamless management of solar, wind, and other distributed energy resources. These capabilities not only enhance sustainability and reduce reliance on fossil fuels but also strengthen grid resilience [65,66,70].

3.3.2. AI Applications for Energy Management in Smart Buildings

AI, both independently and in combination with technologies such as Building Information Management (BIM) systems, plays a pivotal role in improving smart building performance through predictive maintenance, energy optimisation, intelligent climate control, and occupant behaviour analytics. These applications lower operational costs, enhance efficiency, and support global sustainability goals [56,70,71].

Predictive maintenance and fault detection are among the most impactful uses. Machine learning algorithms detect anomalies and failures before they escalate, thereby reducing downtime and costs [58,72]. Advanced Fault Detection and Diagnosis (AFDD) systems set operational baselines, using techniques such as entropy-based analysis to improve diagnostic accuracy. These methods prevent equipment failure and deliver significant energy savings [72].

AI further strengthens energy forecasting and optimisation [68,73]. By analysing historical and sensor data, machine learning, digital twin and edge computing models predict demand trends, supporting proactive energy planning and integration with HVAC, renewables, and demand response [34,37,38,56,74,75]. Predictive control systems can cut heating energy use by up to 20% without sacrificing comfort [73]. When combined with BIM, AI also aids decision-making on materials, cost, energy, and scheduling [76].

In indoor climate control, reinforcement learning enables HVAC systems to optimise performance more effectively than traditional methods, enhancing comfort while reducing energy use [59,77]. AI also monitors air quality, adjusting ventilation to balance efficiency and occupant health [77]. In addition, behavioural analytics enable AI to predict usage patterns and adjust systems, such as lighting and temperature, to match preferences, thereby reducing waste and improving comfort [74,78]. In this context, according to Liu et al. (2023) [79], AI-based controllers and occupant-centric strategies are crucial for achieving high levels of thermal comfort and energy savings in smart, energy-efficient buildings. Their work provides insights into the mechanisms, algorithms, and applications of advanced control systems for smart, energy-efficient (SEE) buildings, offering a pathway toward sustainable, low-carbon transitions in the building sector.

3.4. Synergistic Integration of IoT and AI: State-of-the-Art Applications of AIoT

The convergence of IoT and AI, commonly referred to as the Artificial Intelligence of Things (AIoT), is shaping the development of next-generation smart building ecosystems by enabling real-time decision-making, adaptive control, and seamless interoperability. As the work of Pandiyan et al. (2023) [80] reveals, integrating IoT with AI enhances energy performance in both buildings and urban environments through continuous monitoring and intelligent control. At the grid level, smart systems integrate renewable energy sources while utilising IoT, AI, and blockchain technologies to predict load fluctuations, thereby optimising energy distribution and maintaining stability. At the building level, IoT sensors monitor occupancy, temperature, and equipment performance, while AI algorithms can automatically adjust lighting, HVAC, and other systems to minimise energy waste. Additionally, combining the capabilities of cloud and edge computing enhances energy management by enabling faster data processing and more efficient resource allocation. By combining IoT’s continuous data collection with AI’s predictive and analytical capabilities, AIoT creates intelligent, responsive, and energy-efficient environments that improve both building performance and occupant well-being [81].

AIoT has already shown measurable benefits: intelligent operations can reduce energy use by up to 30% and operating costs by 20% [43]. Analysing sensor data, AI supports predictive maintenance, fault detection, and energy optimisation [30]. Smart systems dynamically adjust HVAC, lighting, and ventilation in response to occupancy and environmental conditions, reducing waste while maintaining comfort [63,81,82]. Semantic frameworks that integrate data sources such as weather forecasts and energy market prices further enhance planning and decision-making [83].

Operationally, IoT sensors monitor structural, environmental, and energy variables, while AI processes this data for adaptive control. Applications include AI-driven HVAC optimisation, intelligent lighting, and autonomous systems that learn occupant preferences [44,84]. Multi-agent systems (MAS) improve coordination among subsystems [85], while edge computing and secure protocols enhance data flow, resilience, and reduce latency [44,86].

However, challenges such as data privacy and security are major concerns given the sensitivity of collected data [14]. Interoperability across heterogeneous IoT devices requires flexible architectures and standards [87]. High implementation costs—often up to 15% of project budgets—alongside technical expertise requirements hinder scalability [43]. Ethical issues also emerge as autonomous decision-making grows, with transparency, accountability, and fairness requiring oversight [88,89].

Future AIoT developments are expected to improve user interaction by providing personalised energy-saving recommendations and real-time monitoring via mobile applications [63]. At the city scale, AIoT will contribute to sustainability and climate goals by reducing building energy demand [14]. Research priorities include secure management of isolated IoT networks, remote power provision, and balancing autonomy with user comfort [68].

Thus, AIoT offers transformative potential for energy management in smart buildings and sustainable urban development. While technical, financial, and ethical challenges persist, ongoing research and cross-disciplinary collaboration will be essential. As it evolves, AIoT is poised to become a cornerstone of intelligent, energy-efficient, and ethically responsible smart cities.

3.5. Barriers to IoT and AI Integration

Despite the benefits of IoT–AI integration in smart building energy management, several barriers hinder large-scale adoption. These fall into three categories: technical, economic and operational, and institutional and social.

3.5.1. Technical Barriers

Key technical challenges include interoperability issues, scalability, and reliance on legacy systems. Interoperability remains a major obstacle as devices from different manufacturers use distinct protocols and data formats. This lack of standardisation complicates integration and introduces security risks [90,91]. Efforts to develop unified protocols and frameworks are essential for building secure, interoperable ecosystems [91].

Scalability is another concern. The rapid growth of connected devices generates massive data volumes requiring real-time analysis [41,92]. Addressing this demand necessitates advanced infrastructure and data management systems [93]. Edge computing offers a promising solution by enabling localised processing, reducing latency, and improving efficiency [92].

Legacy systems further constrain adoption, as they often lack the capacity for modern AI applications. Retrofitting these systems is technically complex and financially prohibitive, especially for resource-limited organisations [94].

Overcoming these barriers requires not only technical innovation but also regulatory support and industry collaboration to ensure scalable, secure, and efficient IoT–AI ecosystems.

3.5.2. Economic and Operational Barriers

High costs, return on investment (ROI) uncertainty, and operational complexity are key economic barriers. Upfront investments in sensors, storage, and processing infrastructure are particularly prohibitive for SMEs [94,95]. Costs are compounded by maintenance, upgrades, and network expansion [96,97,98,99].

ROI uncertainty also slows adoption. Benefits such as efficiency gains and better decision-making often emerge only in the long term, causing many projects to stall in “pilot purgatory” [94]. Unexpected integration challenges and delays further erode returns [100].

Nonetheless, long-term benefits ranging from improved operations to new services can outweigh short-term challenges. Phased implementation, prioritisation of high-impact use cases, and strong leadership support can help organisations overcome financial and operational hurdles [98]. As technologies mature and costs fall, these barriers are expected to decline.

3.5.3. Institutional and Social Barriers

Beyond technical and economic considerations, the successful integration of IoT and AI also depends on overcoming institutional and social challenges. The most critical of these include data privacy and cybersecurity risks, shortages of skilled personnel, and resistance to technological adoption.

Data privacy and cybersecurity are critical concerns. IoT devices generate sensitive data that is vulnerable to breaches, particularly in the absence of standardised security frameworks. Secure protocols, edge computing, and ethical safeguards are essential for protecting integrated systems [86,101,102].

The skills gap is another barrier. The demand for expertise in AI, data analytics, and IoT device management outpaces the supply [42,103]. Resistance to adoption, especially in traditional and public sectors, further impedes progress. Concerns about costs, disruption, and regulatory gaps slow uptake [104]. Demonstrating clear benefits, such as improved efficiency and service delivery, can help overcome such reluctance [105].

Addressing institutional and social barriers will require cross-sector collaboration among academia, industry, and policymakers to expand training programmes, robust security measures, and workforce development [18]. With continued research, policy support, and innovation, IoT–AI integration can deliver intelligent, sustainable, and user-focused building management systems [64,70,77,102].

4. Case Studies

4.1. The Edge, Amsterdam

The Edge demonstrates how IoT and AI can transform an office building into an energy-positive workplace. Thousands of IoT sensors monitor occupancy, temperature, and daylight, while AI-driven analytics optimise HVAC and lighting in real time. This integration enables exceptional efficiency without compromising occupant comfort. The building’s designation as Building Research Establishment Environmental Assessment Method (BREEAM) Outstanding (98.36%) reflects its strong adoption of renewable energy sources and demand response capabilities. Nonetheless, it faces challenges such as high capital costs, interoperability of systems, and cybersecurity risks. Overall, The Edge demonstrates the potential of embedding predictive AI and IoT into large-scale energy management in commercial buildings [43,54,106,107,108].

4.2. Rinascimento III, Rome

Rinascimento III showcases a residential-scale application of IoT, AI, and digital twin technologies. By linking a three-dimensional model of the neighbourhood with IoT devices, the system supports real-time monitoring, predictive demand management, and scenario testing before implementation. With 70% renewable energy integration, the development meets near-zero energy building (NZEB) standards while ensuring resident comfort. Key challenges include high upfront investment, data privacy concerns, and interoperability issues. Despite these, Rinascimento III highlights how predictive analytics and adaptive control can balance sustainability with livability in smart housing developments [43,54,109].

4.3. Infosys Campuses, India

Infosys illustrates enterprise-scale energy management through IoT and AI. Across more than 30 million ft² of campuses, IoT-enabled submeters and AI-powered analytics optimise HVAC, retrofit building systems, and intelligently schedule energy-intensive loads. Between 2008 and 2020, despite a 166% workforce increase, Infosys limited electricity growth to 20%, avoided 2.36 billion kWh of electricity use, and reduced its connected load by 35 MW. The main barriers included retrofitting IoT into legacy systems and upskilling staff in digital management. By institutionalising IoT and AI across multiple sites, Infosys demonstrates how corporations can decouple growth from energy use, making efficiency both a sustainability and business strategy [110].

4.4. Keppel Bay Tower, Singapore

Keppel Bay Tower demonstrates the retrofitting potential of IoT and AI in urban high-rises. As Singapore’s first Super Low Energy (SLE) commercial building, it combined IoT-driven real-time monitoring with AI-enabled chiller plant optimisation, yielding a further 7% reduction in annual energy intensity. A performance digital twin allowed strategies to be tested before deployment, improving reliability and effectiveness. Challenges such as high upfront costs, integration complexity, and reliability risks were addressed through solutions, including auxiliary bypass systems. Recognition by Singapore’s Building and Construction Authority (BCA) underscores the measurable sustainability outcomes achievable through IoT and AI in dense urban contexts [54,111,112,113].

4.5. Lessons from the Case Studies

Collectively, the four case studies demonstrate the transformative potential of integrating IoT and AI in enhancing energy efficiency across a wide range of building typologies and operational contexts. A Comparative Summary of Different Features of the IoT–AI Smart Building Case Studies is presented in

Table 5. Comparative Summary of IoT–AI Smart Building Case Studies.

Case Study	Building Type/Location	IoT–AI Integration Features	Energy Savings/Efficiency Outcomes	Cost and Investment Considerations	Integration Maturity Level	Sustainability/Certification Outcomes
The Edge, Amsterdam	Commercial Office (Netherlands)	Thousands of IoT sensors monitor occupancy, light, and temperature, enabling AI-driven HVAC and lighting optimisation in real-time.	Achieved BREEAM Outstanding (98.36%); significant reduction in operational energy demand (≈30%).	High upfront capital cost; interoperability and cybersecurity remain challenges.	Advanced: Fully integrated AI–IoT platform with predictive analytics and renewable integration.	BREEAM Outstanding; energy-positive design with renewable sources.
Rinascimento III, Rome	Residential NZEB District (Italy)	Digital twin linked with IoT for real-time monitoring, predictive demand management, and scenario testing.	~70% renewable energy integration; meets Near-Zero Energy Building (NZEB) standards.	High initial investment; data privacy and interoperability issues persist.	Intermediate: Strong IoT infrastructure with predictive analytics; partial automation.	NZEB-certified; high renewable share and occupant comfort.
Infosys Campuses, India	Corporate/Multi-site (India)	IoT submetering across >30 million ft2; AI analytics for HVAC scheduling, retrofits, and predictive maintenance.	Electricity use limited to +20% despite 166% workforce growth; avoided 2.36 billion kWh; reduced 35 MW load.	Moderate cost relative to scale; key challenge in retrofitting legacy systems and staff upskilling.	Advanced: Enterprise-wide IoT–AI integration with centralised data analytics.	Corporate sustainability programme; internal net-zero operations targets.
Keppel Bay Tower, Singapore	Commercial High-Rise (Singapore)	IoT real-time monitoring and AI-enabled chiller plant optimisation; performance digital twin for simulation.	An additional 7% reduction in annual energy intensity; an overall Super Low Energy (SLE) rating was achieved.	High upfront and retrofit costs; reliability and integration complexity addressed via adaptive solutions.	Advanced: Mature AI–IoT use with digital twin validation and real-time optimisation.	Singapore BCA Super Low Energy certified; recognised for retrofit innovation.

Note: Integration Maturity Levels: Basic—limited automation or isolated IoT systems; Intermediate—IoT with partial AI analytics; Advanced—fully integrated, predictive, and adaptive AI–IoT ecosystems.

.

A key insight is that the combination of IoT, AI analytics, and digital twin technologies allows predictive, real-time optimisation of building systems. This integration enables a transition from reactive energy management to proactive, data-driven control, improving operational reliability and occupant comfort. The cases of The Edge and Rinascimento III further illustrate that coupling these technologies with on-site renewable energy generation is critical to achieving net-zero or energy-positive performance.

At the organisational scale, Infosys shows how the adoption of IoT-AI systems at an enterprise-wide level can decouple energy use from business growth, demonstrating the scalability of smart energy management as a corporate sustainability strategy. On the other hand, Keppel Bay Tower highlights the potential for retrofitting existing structures with adaptive system integration and performance-based optimisation.

Despite differing contexts, all projects shared some common barriers: high capital costs, interoperability and data governance challenges, and the need for digital upskilling of the workforce. Nevertheless, quantifiable outcomes, such as a 7% annual reduction in energy intensity at Keppel Bay Tower and over 2.36 billion kWh of avoided electricity use at Infosys, point out the measurable benefits of predictive AI algorithms and IoT-enabled monitoring.

Thus, although the route to smart and energy-efficient buildings is complex, context-dependent, and fraught with detours, these cases demonstrate that the integration of IoT–AI offers a robust pathway toward decoupling energy demand from economic growth and helping to advance net-zero objectives while supporting a transition toward sustainable, intelligent, and resilient built environments.

5. Discussion, Strategic Pathways, Integrated Framework, and Policy Implications

5.1. Discussion: Systems Perspective, Barriers, Interconnectedness, and Conceptual Grounding

The convergence of IoT and AI in the built environment represents a pivotal step toward sustainable development. With urbanisation accelerating, buildings account for roughly 40% of global energy consumption, positioning them at the heart of climate mitigation strategies [1]. In this context, IoT provides the infrastructure for sensing and data collection, supporting applications such as smart metering, occupancy-based regulation, and the integration of renewable energy [63,64]. AI complements this infrastructure by functioning as the analytical and decision-making engine, applying predictive models and machine learning techniques to forecast demand, optimise HVAC performance, and automate building operations [59,73]. Together, IoT and AI form a symbiotic system, AIoT, that not only enhances operational efficiency and extends asset lifespans but also reduces emissions and improves indoor environmental quality [8,77].

Yet, despite this transformative potential, adoption remains constrained by a combination of technical, economic, and institutional barriers. Technically, limited interoperability among devices, platforms, and legacy systems often constrains scalability [90,91]. Economically and operationally, high upfront costs coupled with uncertain returns often discourage investment, particularly in resource-constrained contexts [95]. Institutionally and Socially, inadequate data governance, privacy risks, and shortages of skilled professionals further complicate implementation [99,103].

A closer analysis of the identified barriers reveals that economic and operational challenges are the most dominant (frequency = 7), followed closely by technical, institutional, and social challenges (frequency = 6 each). In the economic and operational category, high costs (4) and ROI uncertainty (2) emerged as the principal barriers, reflecting financial constraints and ambiguous returns that impede adoption. Within the technical category, scalability (3) and interoperability among devices (2) were the most frequently cited issues, underscoring persistent difficulties in integration and system expansion. Similarly, institutional and social challenges ranging from data privacy and cybersecurity risks to skill shortages and resistance to technological adoption underscore the human and organisational dimensions of implementing AIoT systems (Supplementary File S5 (Table S1)).

The dominance of economic and operational barriers suggests that the success of AIoT in the built environment depends primarily on improving financial feasibility and economic justification through strategic investments, incentives, and clear return-on-investment models. Furthermore, addressing technical barriers will require standardisation, system interoperability, and scalable architecture, while mitigating institutional and social challenges requires capacity building, data governance frameworks, and change management initiatives to ensure sustainable and inclusive technology adoption.

These barriers seldom act in isolation. Rather, they interact systemically—technical incompatibilities can elevate costs, while weak regulatory frameworks exacerbate privacy concerns—making piecemeal interventions insufficient. Overcoming these challenges necessitates systems thinking, which views buildings not as isolated physical units but as complex socio-technical entities embedded within broader urban and institutional networks [10].

Equally critical is recognising the sustainability of the digital technologies themselves. While IoT and AI optimise resource use, they also introduce challenges associated with energy-intensive data processing, electronic waste, and ethical considerations in algorithmic governance [45]. Addressing these issues requires a cradle-to-grave perspective that encompasses the entire life cycle of digital technologies. Emerging approaches, such as digital twins, federated learning, and distributed energy management, offer promising pathways by facilitating decentralised data processing, improving long-term performance evaluation, and mitigating environmental impacts [33,39]. Consequently, the conceptual foundation for IoT–AI integration lies not only in their enabling role in advancing sustainability but also in understanding their entanglement with broader ethical, ecological, and institutional systems.

5.2. Strategic Pathways to Overcome Barriers

Unlocking the potential of IoT and AI in the built environment requires strategic interventions that address existing barriers in a structured and coordinated manner. The following subsections discuss strategic pathways.

5.2.1. Framework for Staged Integration

A staged integration framework is likely to ensure organisations adopt IoT and AI systematically and in ways that reflect their operational context. The process begins with assessment, in which infrastructure is evaluated, needs are identified, and suitable applications are defined—establishing both feasibility and scope [94]. The implementation phase typically follows, starting with pilot projects that generate feedback, reduce uncertainty, and allow refinement prior to large-scale deployment [94]. Finally, a phase of continuous evaluation ensures that integration delivers intended outcomes while identifying opportunities for improvement and adaptation [94].

5.2.2. Role of Open Standards and Protocols

Successful IoT and AI convergence depends on the adoption of open standards and communication protocols. Systems such as Building Automation and Control Network (BACnet), Konnex (KNX), and Message Queuing Telemetry Transport (MQTT) enable interoperability across devices from different manufacturers, a prerequisite for scalability and flexibility [114]. By reducing integration complexity and promoting compatibility, these standards not only lower costs but also facilitate the seamless deployment of AI across IoT networks [114].

5.2.3. Capacity-Building and Stakeholder Training

Capacity-building and stakeholder engagement are essential to sustainable integration. Developing the technical skills to operate and maintain IoT and AI systems (AIoT) requires comprehensive training programmes [88]. Beyond technical expertise, the active involvement of all stakeholders, from senior leadership to frontline staff, supports cultural readiness for innovation and creates organisational conditions for long-term transformation [94].

5.2.4. Policy and Incentive Mechanisms

Governmental policies and incentives provide critical enablers for adoption. Financial instruments such as tax rebates and green certifications can encourage investment in IoT and AI systems [104]. Supportive regulatory frameworks also address concerns over data privacy, cybersecurity, and ethical AI use. By reducing risks and uncertainty, these mechanisms build confidence among adopters and promote broader diffusion [104].

5.2.5. Cross-Sector Collaboration

Collaboration across academia, industry, and government strengthens innovation capacity and accelerates the adoption of new ideas. Joint research and development initiatives generate advanced solutions and establish best practices [115]. Public–private partnerships, in particular, offer effective mechanisms for pooling resources, sharing expertise, and scaling the deployment of IoT and AI (AIoT) [104].

While these strategic pathways create opportunities, significant challenges remain. Data privacy risks, cybersecurity vulnerabilities, and ethical issues in algorithmic decision-making require careful management through robust security protocols, transparent guidelines, and clear accountability mechanisms. Public perception is also decisive: education and engagement initiatives are necessary to ensure IoT and AI are not only technologically advanced but also trusted, beneficial, and aligned with societal values [116].

5.3. Integrated Framework for Sustainable Energy-Efficient Smart Buildings

The framework emphasises the systemic integration of IoT and AI in smart buildings to drive energy efficiency and sustainability. Grounded in a systems perspective (cf. Section 1.4), it views smart buildings as dynamic networks of interdependent elements—physical structures, technological infrastructure, human occupants, institutional rules, and urban systems—whose interactions shape outcomes. To achieve energy efficiency, the framework is built on three interconnected pillars: positioning IoT and AI as core enablers, addressing integration barriers strategically, and targeting high-impact applications (Figure 4). These pillars are aligned with SDG 7 (Affordable and Clean Energy), SDG 11 (Sustainable Cities and Communities), and SDG 13 (Climate Action).

5.3.1. The Interconnected Pillars

The three interconnected pillars are IoT and AI as Core Enablers, Addressing Integration Barriers Strategically and Targeting High-Impact Application Areas.

Pillar 1: IoT and AI as Core Enablers

The first pillar recognises IoT and AI as foundational technologies for innovation in smart buildings. IoT enables the collection of real-time data through sensors, meters, and devices that monitor energy consumption, occupancy, temperature, and air quality [63,70]. AI builds upon this infrastructure with predictive analytics, optimisation techniques, and autonomous control of building systems such as HVAC, lighting, and security [73]. By working together, IoT and AI create intelligent environments that reduce operational costs, improve comfort, and lower emissions [8]. This synergy is reinforced through renewable energy integration and smart grids, which support distributed generation and storage [65,66]. In addition, digital twins, which are virtual representations of buildings, offer simulation and optimisation capabilities while supporting compliance with sustainability standards [33].

Pillar 2: Addressing Integration Barriers Strategically

The second pillar acknowledges that the transformative potential of IoT and AI can only be realised by addressing technical, economic, and institutional barriers. Technically, interoperability issues and proprietary protocols increase complexity and cybersecurity risks, while retrofitting legacy buildings adds challenges [90,91]. Open standards such as BACnet and MQTT reduce costs and improve scalability [114]. Economically, high upfront costs, uncertain returns, and “pilot purgatory” hinder large-scale adoption [94,95], with further operational expenses for maintenance and data storage. Incentives like tax rebates, subsidies, and public–private partnerships help share risks [104]. Institutional barriers include privacy, data misuse, and algorithmic bias, requiring transparent governance [99]. Skills shortages can be addressed through training and certification [18,103], while participatory design builds trust and support long-term adoption [105].

Pillar 3: Targeting High-Impact Application Areas

The third pillar focuses on high-impact applications where IoT and AI can maximise sustainability outcomes. Predictive maintenance and fault detection reduce energy loss and downtime [73]. AI-based forecasting optimises real-time operations [74]. Advanced climate control and air quality management balance efficiency and occupant well-being through reinforcement and deep learning [77]. Behavioural analytics personalise energy use, reducing waste and improving satisfaction [78]. Aligning these solutions with Leadership in Energy and Environmental Design (LEED), BREEAM, and policy frameworks ensures measurable and verifiable sustainability outcomes.

Integrated Framework

By linking the three pillars—technology enablement, barrier mitigation, and targeted applications, the framework provides a comprehensive, systems-level roadmap for sustainable, energy-efficient smart buildings. It integrates technological innovation with socio-institutional adaptation through continuous feedback loops, in which operational data actively informs policy, design decisions, and optimisation strategies. Key enabling tools include digital twins for performance simulation [33], multi-agent systems (MAS) for coordinating decentralised subsystems [29], and smart energy management systems (SEMS) for forecasting demand and optimising usage [117]. Collectively, these technologies facilitate predictive analytics, adaptive control, and the integration of renewable energy, supporting higher clean energy penetration while maintaining reliability [118].

Evidence from case studies shows that the integration of IoT and AI not only reduces energy use and operational costs but also generates actionable insights for building and urban-scale planning [54,119]. At the urban scale, smart buildings contribute to climate strategies by sharing data with municipal systems and dynamically responding to city-wide energy demands [33,118,120]. However, challenges remain, including interoperability issues, cybersecurity risks, high costs, and skills shortages. Addressing these challenges requires coordinated cross-sector collaboration, supportive regulatory frameworks, and targeted innovation to enable scalable adoption.

The framework positions smart buildings as active contributors to sustainable urban transformation. By combining technological capability, institutional adaptation, and human-centred design, it fosters systemic shifts toward energy efficiency and climate resilience, aligning with global sustainability objectives and highlighting the interplay between technological, social, and policy dimensions.

5.3.2. Operationalising Through Systems Perspectives

The operationalisation of this framework employs systems thinking to map the causal feedback relationships loops (CLDs) and mechanisms through which technology, policy, and behaviour interact. Five key feedback loops—three reinforcing (R) and two balancing (B)—illustrate the processes that amplify benefits or stabilise system performance (Figure 5a–e).

The proposed systems-thinking framework was inductively developed from recurring patterns in the reviewed literature, highlighting interactions between technology, policy, and behaviour in smart buildings and urban sustainability. Key trends—such as IoT and AI-driven performance optimisation, policy incentives, and user responses—were synthesised into five feedback loops (R1–R3, B1–B2), capturing the dynamic mechanisms of sustainable building operations. The framework is grounded in systems theory, integrating empirical evidence with established theoretical principles.

The first reinforcing loop, Operational Performance and Design Improvement (R1), demonstrates that IoT sensors continuously collect operational data, which digital twins and AI algorithms analyse to optimise building systems. These optimisations feed back into building design, improving efficiency and generating higher-quality operational data, thereby creating a continuous cycle of learning and refinement [33] (Figure 5a). The second reinforcing loop, Energy Efficiency and Policy Incentives (R2), operates as reductions in energy use and emissions trigger policy responses, such as subsidies, rebates, or certifications, which in turn increase adoption of IoT and AI technologies. This feedback reinforces energy efficiency gains by linking performance outcomes with external incentives [104] (Figure 5b). The third reinforcing loop, Urban Integration and Resilience (R3), functions through the flow of building-level data into municipal energy and climate systems. Enhanced urban resilience builds trust and encourages further investment in sustainable infrastructure, which strengthens the integration of smart building systems [118] (Figure 5c).

The balancing loops stabilise system behaviour. In the first, Occupant Feedback and System Adaptation (B1), user comfort and behavioural patterns feed into AI-controlled HVAC, lighting, and air quality systems. When discomfort occurs, AI adjusts operations to balance efficiency with satisfaction, maintaining user acceptance [78] (Figure 5d). In Cybersecurity, Trust, and Adoption (B2) mechanism, system vulnerabilities such as data breaches or misuse reduce stakeholder trust, slowing adoption rates. Implementing robust cybersecurity measures and transparent governance counteract these risks, stabilising adoption over time [99] (Figure 5e).

Together, these five loops form an integrated systems framework capturing the interplay of performance optimisation, policy mechanisms, user acceptance, and resilience at building and urban scales (Figure 6). Reinforcing loops (R1–R3) promote adoption through IoT data, policy incentives, and enhance energy efficiency, while balancing loops (B1–B2) balance performance and trust, and prevent instability, ensuring a sustainable, socially acceptable system. Further, cross-loop interactions enhance the framework: occupant comfort informs system adaptation, which fine-tunes efficiency, while policy incentives support adoption and integration with urban energy systems. These interactions emphasise the interdependencies that make the framework resilient and adaptive.

As energy and climate challenges intensify, understanding these mechanisms is crucial for designing and operating the next generation of energy-efficient, intelligent, and sustainable buildings. By linking technological processes, policy instruments, and human aspects, the proposed framework operationalises a coherent system for sustainable transformation. This approach allows smart buildings to dynamically manage resources, enhance energy efficiency, and contribute to urban sustainability.

The framework extends existing smart building and urban sustainability models by integrating technological, policy, and human dimensions within a feedback-loop structure. Unlike existing models—such as AI-Driven Energy Management Systems [73]. IoT-Enabled Building Energy Management Systems [54,121], AI4EF: Artificial Intelligence for Energy Efficiency [122], Human-Centric and Context-Aware IoT Frameworks [123], and Internet of Energy (IoE) Frameworks [124]—which typically address technological or policy dimensions in isolation, this study emphasises the interdependencies among operational performance, energy efficiency, occupant behaviour, and policy incentives. By integrating socio-technical and resilience-oriented perspectives with digital twin–based management, the proposed framework illustrates how data-driven decision-making can dynamically interact with policy mechanisms and user feedback. This holistic integration underscores the framework’s novelty and its practical relevance for advancing both research and applied practice in sustainable smart building management.

5.4. Policy and Practical Implications

5.4.1. Recommendations for Policymakers and Urban Planners

Policymakers and urban planners are central to creating the institutional and regulatory environment required to scale smart, sustainable buildings. Addressing the technical, economic, and institutional barriers identified earlier demands multi-level governance strategies that integrate smart building innovation into broader sustainability agendas. National and municipal policies should mandate or incentivise IoT and AI integration in both new construction and retrofits.

Regulatory bodies should standardise communication protocols (e.g., BACnet, MQTT) and promote open data platforms to enhance interoperability and foster cross-sector innovation [91,114]. Data privacy and cybersecurity safeguards must be embedded in these policies to address public concerns [99]. Governments can further accelerate adoption through fiscal incentives such as tax rebates, certification subsidies, and preferential green loans tied to performance benchmarks [104].

Urban planners should embed digital infrastructure within land-use planning, zoning, and building codes, ensuring smart building design is part of holistic, systems-based development. Such integration aligns with systems thinking by simultaneously addressing energy, mobility, resilience, and equity, thereby advancing the transition to net-zero energy districts.

5.4.2. Guidelines for Developers and Facility Managers

Developers and facility managers operationalise smart building technologies and should adopt phased integration strategies, beginning with pilot projects to evaluate performance and ROI before scaling [94]. Facility managers should prioritise predictive maintenance, demand forecasting, and AI-based climate control to optimise energy efficiency and occupant comfort [73,77].

Capacity building is critical. Developers should invest in cross-disciplinary training for staff in digital infrastructure management, data analytics, and automated systems. Partnerships with technology providers and universities can bridge expertise gaps [103]. Life-cycle cost analysis should guide investment decisions, demonstrating long-term savings. Where feasible, IoT- and AI-enabled systems should integrate with renewable energy (e.g., rooftop PV, storage), optimised through AI to enhance autonomy and reduce grid dependence [33,35].

5.4.3. Implications for Green Building Standards

IoT and AI integration strengthens compliance with international green building standards such as LEED, BREEAM, and Excellence in Design for Greater Efficiencies (EDGE), which increasingly emphasise data-driven performance monitoring and occupant-centric design. Points are awarded for innovations in energy metering, optimisation, and indoor environmental quality—all of which are enabled by IoT and AI [25].

Smart technologies also support commissioning and post-occupancy evaluation, facilitating long-term performance verification. Certification bodies could further advance innovation by incorporating digital readiness criteria, such as interoperability, cybersecurity, and AI ethics, into their frameworks. This would encourage developers to treat intelligent systems not as optional add-ons, but as essential features of sustainable design.

5.4.4. Contribution to Global Climate Targets and the SDGs

Smart buildings enabled by IoT and AI contribute directly to SDG 7 (Affordable and Clean Energy) through improved energy efficiency and renewable integration. They reduce emissions, enhance infrastructure resilience, and improve quality of life, supporting SDG 11 (Sustainable Cities and Communities). By optimising resource use and minimising waste, they advance SDG 12 (Responsible Consumption and Production), and through greenhouse gas reductions, they contribute to SDG 13 (Climate Action). These impacts align with the Paris Agreement’s net-zero targets and enhance the case for climate finance and green investment, positioning energy-efficient smart buildings as catalysts for mobilising low-carbon resources.

5.4.5. Ethical and Environmental Considerations

Aside from technical and operational factors, the development and implementation of IoT- and AI-powered smart buildings also raises relevant ethical and environmental concerns. The protection of data privacy and cybersecurity will be vital in terms of sustaining the trust of occupants and making deployments socially responsible [93,99]. The lifecycle impacts of sensors and associated digital infrastructure, including energy use, e-waste, and material consumption, must be considered to ensure that the overall sustainability outcomes of AIoT systems are net positive. Furthermore, embedded algorithmic transparency and accountability will be crucial to ensure that unintended inequities or suboptimal energy management do not persist due to a lack of clarity in the AI-driven decision-making process. Embedding these considerations into governance frameworks, design standards, and operational protocols would enhance the environmental, social, and ethical value of smart buildings, thereby better aligning their deployment with global discussions on responsible AI and sustainable urban infrastructure.

Thus, advancing smart, sustainable, energy-efficient buildings requires a systemic, multi-stakeholder approach that unites policymakers, planners, developers, and certification bodies. Coordinated action—through incentives, technical standards, updated certification frameworks, and integrated planning—ensures that IoT and AI not only automate building functions and enhance energy efficiency but also catalyse the transition to low-carbon, equitable, and resilient urban futures.

6. Conclusions

The study examines the integration of IoT and AI as transformative enablers of sustainable, energy-efficient smart buildings. Using a systems-thinking perspective, it evaluates the potential, applications, and adoption barriers of these technologies for building energy efficiency. It proposes a framework based on three interdependent pillars: establishing IoT and AI as foundational infrastructures, addressing integration barriers through strategic interventions, and deploying high-impact applications to maximise energy efficiency and sustainability outcomes. Drawing on contemporary literature and case studies, the study shows how IoT–AI-enabled systems can reduce energy consumption, enhance indoor environmental quality, lower operational costs, and advance broader environmental and social objectives. The study identifies technical, economic, and institutional barriers to adoption and proposes strategies such as open communication protocols, supportive policy incentives, investment in capacity building, and stakeholder engagement. These interventions provide a roadmap for scalable implementation across diverse urban contexts.

The findings also reaffirm that IoT- and AI-enabled building systems are not simply technological upgrades but essential pathways for sustainable urban transformation. By aligning with global sustainability frameworks—particularly SDG 7, SDG 11, SDG 12, and SDG 13, smart buildings emerge as vital tools for urban decarbonisation and resilience. Specifically, the proposed framework offers a practical roadmap for policymakers and practitioners by linking technological innovation with governance and capacity-building, showing how IoT–AI integration can be operationalised to accelerate net-zero building transitions and data-driven urban policy.

Unlike many prior models, which have often considered technology or policy in isolation, this framework explicitly integrates the technological, policy, and human dimensions within a feedback-loop structure. It operationalises digital twin-based management, embeds occupant behaviour, and models reinforcing and balancing dynamics through five feedback loops for more adaptive and resilient system perspectives at building and urban scales. As a result, this uniquely bridges technical design and policy implementation on real-world adoption, both at the building and city levels.

This study contributes conceptually by positioning IoT and AI as embedded systems rather than optional add-ons, highlighting their role in shaping sustainable outcomes through the interaction of technological capabilities, infrastructural constraints, and policy mechanisms. Practically, it synthesises applications such as predictive maintenance, intelligent climate control, and occupancy-based energy management, and outlines strategies to address challenges including interoperability, costs, and data privacy. At the policy level, it offers actionable recommendations for governments, urban planners, and facility managers, emphasising open standards, financial incentives, capacity building, and regulatory clarity. By linking smart building innovation to global agendas, it positions intelligent infrastructure as central to advancing climate action, energy efficiency, and urban resilience.

The study has certain limitations. It is based primarily on qualitative synthesis, requiring validation through longitudinal real-world implementations. The rapid evolution of digital technologies may render some solutions obsolete, underscoring the need for adaptive governance. Moreover, the environmental costs of producing and disposing of smart devices—including embodied carbon and electronic waste—remain underexplored. The literature reviewed also shows a regional bias toward the Global North, with limited evidence from the Global South. This might restrict the generalisability of the framework, and thus it should be adapted to different socio-economic, climatic, and policy contexts, especially in developing regions where infrastructural and digital capacity remains limited.

Future research should focus on empirical validation through pilot projects and post-occupancy evaluation of AIoT-enabled buildings, complemented by life-cycle assessments to account for the environmental footprint of devices. Studies could also focus on the role of IoT and AI in retrofitting existing buildings and in promoting lifestyle-driven energy efficiency, providing practical strategies for both legacy infrastructure and occupant-centred sustainability. The intersection of digital ethics, sustainability, data governance, surveillance, and equity also deserves attention. More localised research is required to position the framework in varied urban typologies and capacities, particularly in the Global South. With regulatory systems still evolving, further work is needed on digital policy frameworks and governance models to ensure the secure, transparent, and inclusive deployment of smart building technologies at scale.