Construction AI

Three industry scenarios

 

McKinsey’s Artificial Intelligence: Construction Technology’s Next Frontier (Agarwal et al 2018) is one of a series of recent papers from the management consultants on AI, automation and infrastructure. They identify five AI-powered applications, and use cases that have already arrived in other industries, that can be applied to construction. This is a practical approach that seems to target major contractors, and is a different approach to previous reports that could have been primarily intended for public sector clients. McKinsey has been seriously developing their infrastructure practice for some years now, positioning themselves for the global infrastructure boom they forecast over the next few decades. The five industry applications are:

•  Transportation route optimization algorithms for project planning optimization;
•  Pharmaceutical outcomes prediction for constructability issues
•  Retail supply chain optimization for materials and inventory management
•   Robotics for modular or prefabrication construction and 3-D printing;
•  Healthcare image recognition for risk and safety management.

Each of these has a short discussion with some examples of crossover potential. They are all plausible extensions of current technology, and in robotics, 3-D printing and drones, leading construction firms are already well advanced. Using AI for optimization is obvious (Gans 2018), and is addressed below (see figure 2), but construction firms typically contract out specialized tasks such as design and logistics, rather than invest in the hardware and software development needed (Manly and XXX). Its questionable whether McKinsey makes a convincing case for using AI in construction. Are these are the pathways into construction for AI, or the only ones?

McKinsey also looks at some machine learning algorithms that are relevant to contractors, and briefly assesses their potential engineering and construction applications. Despite their extensive reporting on BIM elsewhere there is no discussion of the potential use of AI in design and engineering, or in restructuring processes. They do have a generic framework for types of machine learning, and they suggest algorithms will be useful for: refining quality control and claims management; increasing talent retention and development; boosting project monitoring and risk management; and constant design optimization.

If McKinsey has a more nuanced story to tell on pathways for AI into construction it might look something like the scenarios depicted by the World Economic Forum and the Boston Consulting Group in their Future Scenarios and Implications for the Construction Industry (WEF/BCG 2018). This scenario analysis is the second, final step in their Future of Construction project, which has involved people from industry and researchers from a wide range of organizations, after the Shaping the Future of Construction report (WEF/BCG 2016). They use infrastructure and urban development Industry (IU) to describe what has elsewhere been called the built environment sector.

The three future scenarios the WEF describe make technological context central to the future form of the industry. The scenarios depict three extreme yet plausible versions of the future. Each scenario is used to extrapolate implications for the industry, identifying potential winners from technological transformation, and the range of examples and ideas shows the value of such a widespread collaboration between industry, government and academia. The WEF does not say how far into the future they are looking, although it seem to be a lot further than McKinsey:

1.       In Building in a virtual world, virtual reality touches all aspects of life, and intelligent systems and robots run the construction industry. Interconnected intelligent systems and robots run the IU, software players will gain power, and new businesses will emerge around data and services.

2.       In Factories run the world, a corporate-dominated society uses prefabrication and modularization to create cost-efficient structures. The entire IU value chain adopts prefabrication, lean processes and mass customization, with suppliers benefiting the most from the transition and take advantage of new business opportunities through integrated system offerings and logistics requirements.

3.       In A green reboot, a world addressing scarce natural resources and climate change rebuilds using eco-friendly construction methods and sustainable materials. Innovative technologies, new materials and sensor-based surveillance ensure low environmental impacts, so players with deep knowledge of materials and local brownfield portfolios thrive on the new business opportunities around environmental-focused services and material recycling.

It is important to keep in mind that scenarios are not predictions of the future. Rather, they outline a broad spectrum of possible futures. In the real future, the construction industry will most probably include elements of all three, as the supply side of changes in demand for different types of building.

One issue is where the industry is at in regard to technology take-up, now that there is widespread recognition of the reality of a digital future. Will construction industry development over the next decades absorb the impacts of new technology and be gradual, changing industry practice over time without significantly affecting industry structure or dynamics? Given the entanglement of economic, social, political, and legal factors in the construction technological system this might be the case, however there are good reasons to think this may be wrong. Machine learning, AI, automation and robotics are an interconnected set of technologies that are evolving quickly, enabled by expanding connectivity and the massively scaleable hardware available today.

 

In 2016 a scenario analysis called Farsight for Construction, looking at the future of the building and construction industry in Queensland, Australia, was released (Quezada et al, 2016). The scenarios describe “four plausible futures for Queensland’s construction industry over the coming two decades, with a focus on impacts for jobs and skills. Each scenario consists of a description of Queensland’s construction industry in the year 2036, a narrative of how the scenario came about, and a commentary on plausibility.” In the figure below Australia is substituted for Queensland.

Figure 1.

Source: Quezada et al, 2016.

If we think of the structure of the construction industry as a pyramid of different sized firms, there is a broad base of tradesmen and small firms at the bottom, followed by a deep layer of medium sized firms, and a small top triangle with a few large firms. Some of those large firms, and some of their major clients, are clearly on the technological frontier, and their investment in capability and capacity should deliver significant increases in efficiency and productivity, and probably scale. Some medium-size firms are also making these investments, and also have access to technologies like algorithmic optimisation, platform-based project management, robotic, VR and AR applications and so on. The WEF Shaping the Future of Construction report (WEF/BCG 2016) included snapshots of what a range of firms at the frontier were doing. These examples reflect the diversity of the industry, and were missing from McKinsey’s high level analysis.

A period of technology-driven restructuring of the building and construction industry may be about to start, similar to the second half of the 1800s when the new materials of glass, steel and reinforced concrete arrived, which led to new methods of production, organisation and management. There are many implications of such a restructuring. Some firms are rethinking their processes in response to developments in AI, robotics and automation as capabilities improve quickly and the range of new products using these technologies expands. Many firms, however, are not. Meanwhile, firms at the frontier are exploring new technology and pushing the boundaries of what is possible, and are inventing new processes.

 

References

Agarwal, R., Chandrasekaran, S. and Sridhar, S. 2016. Artificial Intelligence: Construction Technology’s Next Frontier, McKinsey & Co.

Quezada, G., Bratanova, A., Boughen, N. and Hajowicz, S. 2016. Farsight for Construction: Exploratory scenarios for Queensland’s construction industry to 2036, CSIRO, Australia.

WEF/BCG, 2016. Shaping the Future of Construction: A Breakthrough in Mindset and Technology, World Economic Forum and the Boston Consulting Group, Geneva.

WEF/BCG, 2017. Future Scenarios and Implications for the Construction Industry, World Economic Forum and the Boston Consulting Group, Geneva.

Recent Books on Construction Economics

Developments in Construction Economics Research

Issues around the measurement, structure and performance of the building and construction industry, and its relationship with the economy and manufacturing, professional services and materials industries, have become the focus of CE in a series of books published since 2008. The contributions developed topics identified within the scope of CE in previous work, but they ranged widely and have consolidated the boundaries of CE while continuing to introduce ideas from elsewhere in economics. For example, the six books have contributions on the activities of large, international contractors that dominate the global construction industry, a topic that has been, and is, of continuing interest.  However, the global perspective in these books, while not new, marked another expansion of the topics and issues addressed, to include developments in market analysis, contractor strategies and in particular international cost comparisons and construction data. The following summary of the topics covered in the books illustrates the scope of CE research and current areas of interest.

The first two books ranged across practical, empirical and theoretical topics. In Economics for the Modern Built Environment (Ruddock 2008) seven contributions were on macroeconomic topics such as the economic effects of capital formation and investment, using construction statistics. There were five studies of markets and contractors, with three contributors emphasising the increasing divergence between global firms and local markets and two country case studies. The book brought a great deal of data together, and updated previous work in empirical CE on measuring construction activity and the broad construction industry. Modern Construction Economics: Theory and application (de Valence 2011) took an industry economics/industrial organization approach with contributions on market structure and competition, auctions and innovation. There were two on production theory and three others on methodology and experimental methods. Three of the contributions directly attacked the model of perfectly competitive markets with price taking firms, arguing construction markets can be concentrated and oligopolistic.

Between them the two books covered many topics and techniques and they carried on earlier debates over production theory and methodology. They included global and national research using macroeconomics, research based on industry economics, and case studies with managerial economics.  Importantly, they consolidated the expansion of the focus of CE from the SIC construction industry and its activity and management, and made the case for CE being about the economics of the built environment. Bridging the gap between the urban scale of the built environment and new building and construction projects, which will typically only deliver a few percent of the total stock each year, has always been a fundamental challenge for CE.

In Measuring Construction: Prices, output and productivity, Best and Meikle (2015) put the focus on data quality and international comparisons of construction costs, raising issues in the collection and use of construction data. As their introduction makes clear “there are standard methods for measurement of physical building work, but the same cannot be said for the characteristics of the construction industry” (p. 1). The twelve contributions covered measurement of construction work, productivity and prices at the global, national, industry and project levels. Their conclusion was “there is no ‘correct’ answer to any of the questions this book explores … It is perhaps only by applying a variety of techniques to the various problems and comparing the results that we obtain that we will know if we are getting closer to developing an acceptable set of tools and methods.’ (p. 256). A multiple models approach is indeed required to tackle the ‘various problems’ with construction data.

In Accounting for Construction: Frameworks, productivity, cost and performance (Best and Meikle 2018), the dozen contributions looked at different ways of measuring construction. With chapters on construction statistics, productivity, costs and data, the book both reviewed and extended previous studies. An ‘important thread’ in the book was “the lack of consistency in the way construction industry data is collected and how it is aggregated and/or disaggregated” (p. xiii). This thread became the focus of the next book in the series, Global Construction Data (Gruneberg 2019). The ten contributions included three on construction statistics, four used cost data, and the other three covered innovation, architectural services and international contractors’ make-buy decisions. In the title the book made explicit this important agenda in CE research. The reliability and quality of construction statistics is a well-known issue, going back to the 1960s, and the shortcomings of the SNA and SIC have not been overcome in the revisions since then. Those shortcomings are also a major theme in Best and Meikle (2015, 2018).

The sixth book in the series is Gruneberg and Francis’ The Economics of Construction (2019). They provide “a game theory account of the behaviour of firms”, the approach typically taken in other branches of industry economics. They discuss aspects of firms’ business models, financing, contractual disputes and power relations at length. A feature is the use of case studies of the collapse of UK contractor Carillion in 2018, Grenfell Tower, construction for the London Olympics and manufactured housing, demonstrating how the business environment a construction firm faces has become significantly more complex over the decades. The profit maximizing firm has evolved into one primarily concerned with growth and survival.

This work continues. The Elgar Companion to CE Research and volume 3 in the measuring construction series, again edited by Best and Meikle, are due in 2021.

 

References

Best, R. and Meikle, J. (eds.) 2019. Accounting for Construction: Frameworks, productivity, cost and performance, London: Taylor & Francis.

Best, R, and Meikle, J. (eds.) 2015. Measuring Construction: Prices, Output and Productivity Abingdon: Routledge.

de Valence, G. 2011. (ed.). Modern Construction Economics, London, Taylor and Francis.

Gruneberg, S. (ed.) 2019. Global Construction Data, London: Taylor & Francis,

Gruneberg, S. and Francis, N. 2019. The Economics of Construction,

Ruddock, L. (ed.) 2008. Economics for the Modern Built Environment, London: Taylor and Francis.