The construction industry presents a vast array of opportunities to innovate through the application of Machine Learning.
From smart assistants that can analyse and prioritise the mountain of data generated throughout the course of a project (as well as identify hazardous and unsafe practices), through to the use of 3D modelling and inventory control, Machine Learning (ML) has the capability to dramatically transform the way the construction industry operates.
Not only will it make construction more efficient and innovative, the ability to capture and analyse project trends will also vastly improve the safety of construction sites.
The Digitisation Imperative
In order for the industry to take advantage of machine learning, there is still much work to do to facilitate this transformation. Getting the basics right is key. This means digitising day-to-day processes, as you are then getting data onto a centralised system, which in turn provides the opportunity to acquire and utilise Machine Learning capabilities.
You can’t utilise Machine Learning until you have data. As more and more businesses adopt construction management software, construction firms are collecting a lot more structured data, opening the door for Machine Learning to be applied to new and existing processes.
Given the significance of the construction industry as a proportion of the Australian workforce, accounting for around nine percent of the total (almost 1.2 million people), the opportunity for large-scale transformation is evident.
Leveraging ML technology to evaluate job progress, the performance of workers, and the availability of materials, will result in projects being completed quicker, to a higher standard, and at a much lower cost. It is clear that this is going to be a game changer for the industry.
Raising the Bar on Safety
Safety is a huge priority in the Australian construction industry. In Procore’s recent Safety InSite report, more than half the businesses surveyed (57%) said they believed that new technologies will help improve safety within their business. Machine learning tools are at the cutting edge of this transformation.
For example, being able to perform tasks such as identifying and tagging safety hazards based on visual data generated on-site (photographs), and identifying and prioritising safety concerns and issues throughout the life of a project, will take workplace safety to a whole new level.
Furthermore, onsite project data will help project managers swiftly identify teams and individuals partaking in unsafe practices, so that they can either be issued with a warning or removed from the job before their actions result in an injury, or worse.
In the current climate, ML can also help companies monitor the necessary safety protocols introduced to inhibit the spread of COVID-19. For example, it can automatically scan photos and videos from job sites to determine whether the social distancing rules are being adhered to.
Streamlining Everyday Tasks
Machine Learning uses data from previous actions and processes to predict what to do next. Think about it in terms of a Google search. Often, Google will complete your search request for you before you have time to type it out fully, based on what you have searched for previously.
When applied to the construction industry, ML speeds up processes through actions such as recommending the right forms to use, or monitoring inventory levels and prompting action when supplies become low, in order to avoid delays due to sites running out of materials and having to wait on new deliveries.
It can also be used to flag issues that might lead to on-site delays, based on what has happened previously in such circumstances.
Tools such as SmartVid leverage a database of site photos and videos, and apply ML to help predict accidents by identifying potential hazards and safety blind spots. Being able to aggregate visual and audio content from multiple formats and pull out relevant safety-related data drastically speeds up what would be an otherwise lengthy and onerous process. It also eliminates the risk of any safety red flags going undetected.
Being able to identify individuals who aren’t following on-site protocol, such as not wearing the appropriate PPE (hard hats, etc.) and automatically generate reports to site managers who can then act to correct the issue, is another very valuable use of ML.
ML-equipped voice assistants can also help site managers and other team leaders perform searches and find the right information quickly. Natural Language Processing, whereby computers are equipped through ML to understand the natural language used by individuals to perform commands, enables users to ask direct questions and get direct answers, rather than having to use very specific language in order to ensure the correct information in response.
Prevention is Better Than Cure
By offering insight and establishing connections between information and safety, Machine Learning would be an asset to the construction industry on that basis alone, but its benefits go way beyond that.
ML engines can analyse the jobsite schedule and determine which tasks require a certified operator of a specific piece of equipment on a certain day, then establish whether that person is due to be onsite that day. Flagging this sort of information helps ensure a delay doesn’t occur through not having an appropriately qualified person available.
This technology can also identify patterns in on-site injuries (where they are occurring and what type of injuries are most prevalent) and ascertain what the contributing factors were, to reduce the risk of further injuries arising. For example, high risk tasks could be flagged, automatically generating a message to all staff to take extra care when completing those tasks.
Ongoing monitoring of the progress of works can also identify lapses in scheduled work taking place, as well as ensuring that standards of workmanship are met at all times.
All of these factors are strong enough in isolation, but combined, they make the case for ML adoption in construction utterly compelling. The first step is digitising day to day processes, and from there, the possibilities are endless.