Data analytics and AI in the construction industry – bridging myths and reality

Digitisation in construction: the current landscape

Over the last few decades, digitisation has left its mark on the building and construction industry, shaping its various facets at different paces. The journey began with the utilisation of digital tools for design management, expanded into project management, and has more recently begun to influence site management.

The incorporation of digital solutions in the industry commenced during the 1980s, with the availability of affordable desktop computers. This marked the start of digitising 2D drawings. Over the subsequent two decades, this transformation progressed further with the democratisation of 3D design. Leading software vendors in this arena include Autodesk and Bentley. While 3D models notably improved design coordination, their overall impact on project productivity has arguably remained limited.

The proliferation of the internet and the development of online project management tools catalysed the digitisation of project management. Pioneering companies such as Procore, Oracle, and Aconex played significant roles in this shift. These tools excel in managing contracts, costs, and schedules. They allow companies to better manage contractual requirements and facilitate controls on large and complex projects. However, their influence on overall productivity remains relatively modest, as project delays and budget overruns persist despite the adoption of these advanced tools.

The digitisation of construction site management, the last aspect to be influenced by digital tools, fully entered the scene with the widespread use of mobile devices about a decade ago. While numerous companies offer solutions in this realm, a clear global market leader has yet to emerge. This third wave of digitisation has the potential for the most profound impact on the industry. Multiple studies emphasise that standardisation and automation of field processes are crucial for boosting productivity, with digitisation being a pivotal component for achieving these goals. However, the construction industry’s inherent nature — centred around individual projects — is a challenge for implementing standardisation and automation.

In spite of these challenges, recent years have witnessed significant progress in construction site management. The focus primarily revolves around three core areas: quality assurance, health, safety, and environmental (HSE) compliance, and resource tracking and allocation. Within these domains, companies are automating and digitising repetitive processes, such as safety compliance procedures, quality controls, punch lists or site diaries. The driving force behind this adoption is increased efficiency and productivity.

The acceleration of digitisation has led construction companies to amass vast amounts of data, prompting the natural progression towards data analysis.

Improving health and safety on sites

HSE field processes are often repetitive and are frequently among the first to be digitised in the field. This includes applications for permits to work (PTWs), toolbox meetings, safety inspections, incidents, and near-miss reports, among other things.

Over the past decade, construction firms have focused on automating and digitising these processes, along with creating dashboards and key leading indicators. These advancements have helped on-site teams reduce compliance management time and streamline reporting efforts. With the growing accumulation of data in the field, the next natural step is to consider whether predictive models could help companies identify unsafe practices.

For companies that have gathered years of accurate and reliable data, the emergence of models capable of predicting high-risk activities is becoming evident. Interestingly, the most valuable data collected often relates not to descriptions but to behaviours. Consider this example: analysing responses in permits to work applications usually does not help predict potential risks or issues. However, the way permits to work are submitted, approved, and patterns in rejection rates offer much more insight into risk patterns.

And this is just the beginning. The collection of field data is set to accelerate further with the integration of video cameras and field sensors. Combining this diverse range of data will undoubtedly improve predictive models. It becomes more and more apparent that Health, Safety, and Environmental (HSE) concerns are where AI will have the most significant impact on the industry in the short term.

Enhancing quality assurance

Quality assurance in construction plays a pivotal role in ensuring projects meet established standards and adhere to specified requirements. One of the initial aspects to undergo digitisation a few years ago was inspections and defect management. These processes, due to their repetitive and straightforward nature, adapted well to digitisation. Within this domain, analytics prove to be very useful on large projects where managing thousands of rectifications is a challenge. Analytics contribute to tracking progress, categorising defect types, and assessing subcontractor performance.

For clients who have standardised and digitised rectification processes with well-defined procedures, machine learning algorithms can accurately offer insights into rectification timelines, subcontractor performance, and in some cases, accurately predict future issues.

Improving defect rectification management is good but it does not tackle the source of the quality issues. The road to improved quality relies on establishing rigorous procedures to pre-empt or promptly identify issues and defects. This is closely linked with Inspection and Test Plans (ITPs). Construction ITPs not only facilitate superior quality work by swiftly identifying and managing issues but also guarantee compliance with ISO 9001 standards—the international benchmark for construction quality management.

Digitising ITPs reaps multiple advantages. It streamlines site operations, reduces rework, and augments efficiency. Forward-thinking construction enterprises are moving away from paper-based reports and forms. This shift is more complex and only a few companies possess abundant, dependable data but we anticipate swift progress in this arena in the forthcoming years, particularly analysing historical ITP data to discern recurring issues and vulnerable, defect-prone areas.

Streamlining progress tracking and resource management

The first things to go digital on construction sites in this area were site diaries, using mobile phones to capture data. Digital site diaries can be simple — just photos and notes — or more advanced and include inputs on manhours, equipment usage and supplies consumption.

When working on large-scale projects, the ability to aggregate data from different sources becomes important. Digital site diaries help project managers gather information from different teams and subcontractors, providing a comprehensive overview of the project’s progress. The info collected can help identify why productivity goes up and down and gain insights into a team’s efficiency, resource allocation, and equipment usage. With accurate ratios from past projects, it also becomes possible to create more precise bids.

With so much data gathered in the field, applying machine learning models to predict risks and delays on projects is a logical next step. However, in most cases, this proves to be a challenge and results are not conclusive. These models can identify risks or issues, but not necessarily more effectively than humans. The challenge here is that variations from project to project are often too significant to generate accurate insights. While these models will improve over time, it will take a few more years and more innovative ways in collecting accurate data before we see a significant impact of AI in this area.

Nevertheless, when combined with dashboards, digital site diaries provide project teams with real-time information that is highly useful.

The future: Integrating mobile apps, IoT and AI

Let’s recap. After over a decade of digitisation and automation in field management, construction firms are witnessing increased productivity in compliance management, quality assurance, project tracking, and resource allocation. Data analytics at the project level is becoming a reality, and leading firms are beginning to utilise machine learning algorithms for insights and informed decision-making.

However, a word of caution regarding AI: all these models depend on large amounts of field-captured data. The main challenge here is not the machine learning models — technology already exists — but rather companies consistently collecting accurate and reliable field data. This process takes time and there’s no shortcut. Nonetheless, the entire industry is steadily progressing in that direction.

So, what comes next? What is the upcoming digital trend in construction? We predict an accelerated digitisation of site management with ever-improving mobile apps that increasingly match the way companies operate in the field, refined AI models, and additional data gathered from IoT systems. This includes devices like video cameras, field sensors, and digital watches. Clients who adopted these solutions in recent years often faced disappointment—these devices were not designed for tough environments, or they were too expensive. Yet, hardware makers are making substantial strides, and in addition, merging IoT data with mobile app data to generate excellent datasets.

And notably, in this domain, we see authorities in Singapore and Hong Kong adapting laws to encourage the use of such systems, particularly for compliance management. It’s conceivable that a wave of innovations may emerge from these countries in the years ahead, accelerating further the fusion of mobile apps, IoT, and AI and accelerating a new reality on sites – a reality where compliance, quality assurance, and project tracking thrive.