How AI could add $5 billion to Kiwi construction by 2035

By david carpenter
Project Director

One of the most important tools to boost productivity and lift our economy to the next level, particularly in the construction sector, is artificial intelligence.

AI, machine learning and metadata are no longer the stuff of sci-fi. Artificial intelligence is becoming an everyday application: each time you shop on Amazon, log onto Facebook or walk into a supermarket past the cameras, you’re interacting with artificial intelligence. AI contributed $2 trillion to global GDP last year, according to one estimate by PwC and Global Investors, and by 2030, that is forecast to reach $15 trillion as the technology advances. Locally, MBIE and the AI Forum estimates that through “labour conversion alone, AI has the potential to increase New Zealand GDP by up to $54 billion by 2035.” Air New Zealand, Xero, Inland Revenue and the New Zealand Defence Force are all using AI to improve processes and safety for their workers.

The potential to address our productivity problem
In the construction sector, we’re still a long way off making data work for us. AI is currently under-utilised in the New Zealand market both onsite and within the project processes. The reluctance to embrace the potential of AI for building projects is understandable.  Consider the enormous amount of information that is produced by a large construction project; it’s almost overwhelming. We tend to deal with data section by section, divided into silos for separate teams, often failing to communicate across the whole project. This is a massive waste of potential. The economic impact of AI to the construction sector is projected to be between $2.4 billion and $5.8 billion by 2035 – if we fail to embrace AI, $3 billion might be wiped off the potential benefit to our industry.

So how can clients use AI to navigate the expansive world of data available to them and capture that which is relevant to their future? 

The first opportunity is to take the information out of an organisation’s silos and centrally use it to improve efficiency and connectivity around a building or infrastructure project. On-site mobile tracking apps, building information modelling, point cloud surveying, risk logging, issue logging and build registers, to name a few of the sources of that data. Decision-makers and contractors can have access to real-time information allowing them to plan and manage projects, but many are further and further removed from influencing the outcomes of projects. Once that data is being accurately captured, AI can be applied to make better use of it.   

What AI on your project site might look like
What could that look like on a major construction project? Take at some of the most advanced technology being used around the world today and think about how it might look on a site in New Zealand if we put it all together. Say Tai is an electrician working on a multi-building commercial project called The Interchange. When he arrives at The Interchange, he signs in via his tablet, reading the site risks and also task-specific risks generated by AI: it’s raining so he’ll need to clean up some water and put down the new anti-slip mats that were ordered to site by the system when stocks got low. As he signs in he picks up his site clip, which identifies Tai’s location and the number of other workers on site. The clip also has a built-in gyroscope that alerts the site manager if he trips or falls, and an alert button sends a signal should he be injured.

Tai notices that there is an amended design to the works he is completing today, so he heads to his works site via the site office to pick up the amended drawing. He now knows, thanks to his tablet, that if he walks to the north side of the site to get to the site managers office, he will be 12% more time efficient than if he went the more obvious route. These efficiencies are adding up to make this contractor’s team 10% more productive than other site contractor teams, which is a contributing factor to the early projected completion date. Previously the team was only 30% productive when on site; AI analysis of the data has helped raise productivity to over 40%. 

Tai picks up the new design from the site manager, Tim, who is heading out on site with the architect, Samira. Tim and Samira each take HoloLens wearable headsets, which use augmented reality to project a model of the eventual build onto the site. This means they can clearly see the head-heights available once the HVAC system is fully installed, one example of the kind of clash detection technology that is helping iron out problems before they crop up.   

Tai is able to finish a job today now that the scissor lift is available – it wasn’t available last time he was onsite so completion was delayed. He marks the task as completed on his tablet and it pops up the ‘Reason for delay’ page with ‘Plant shortage’ at the top of the list; he ticks this box and the ‘Percentage complete’ for the whole project repopulates to include his contribution. 

Roger, the Programme Director, is about to head to the steering group meeting. He has been monitoring plant shortage bottlenecks on a number of projects and found that in this year alone it has accounted for 8% of total delays. He’s come up with a solution that could reduce plant shortages, which will equate to huge savings – and he’s just seen another notification of a task delay due to plant shortage from an electrician on site, which strengthens his case for making improvements

From risk mitigation to change orders to autonomous diggers, AI can transform every part of the construction process. Communication is probably the easiest starting point and one that will make an enormous difference – project success is always aligned directly with communication and collaboration from the senior management to the contractor.

New Zealand not only needs to embrace AI, we need to catch up on the use of more basic types of technology:  virtual meeting rooms, smartphone apps with real-time updates and one-click reporting. Managers can now be agile around correcting bottlenecks on the project. The principle is far-reaching, yet we often see important key information – such as issues logs, lessons learnt, best practice and issues resolution – languishing in archives and never even analysed or used to anywhere near its potential. For us to catch up to other nations when it comes to AI, we’re also going to need to catch up on broader innovation and use of data.

Barriers to AI adoption
Why aren’t we adopting more AI, more quickly? It’s a combination of factors, with the biggest barriers being the time, effort and upfront cost of making the transition to an integrated centralised system.  We need to think carefully in the construction sector where the investment should lie, which will vary from project to project.

A fear of the unknown and new technology is another barrier, along with disrupting our current ways of working. Data logging is also seen as additional, or even pointless work, when a clear overall picture is not defined. And, in the end, while the technology is still evolving and has not been adopted consistently across the construction sector, it’s hard to see best practice models and the inherent cost of being an early adopter is tough to overcome.