The property, construction and infrastructure sectors are beset by increasing complexity and risk, while generating a rising tide of data. At the same time, advanced digital tools and artificial intelligence are growing in power and sophistication and transforming decision making. Bringing these worlds together is a must – but construction professionals have been slow to innovate and adopt.
For developers, asset owners and construction professionals, keeping up with the evolution of technology may seem daunting, but it’s clear how much we stand to gain: faster, more informed decision-making, reduced risk exposure, and increased confidence in delivery outcomes.
The emergence of ‘agentic AI’ is the latest chapter in this story, so let’s take a look at what it can offer.
The new frontier: ‘agentic AI’
‘Agentic AI’ is the next phase in the evolution of what most people are already reasonably familiar with, such as predictive analytics, machine learning and large language models – like ChatGPT. These simpler forms are often called ‘generative AI’. Ask a question, and the model gives you an answer based on the information it can access and synthesise. It is only as good as its training and the data it can access.
‘Agentic AI’, also known as ‘multi-agent orchestration’, represents a major uptick in generative AI’s scale and power – but the premise is still reasonably simple (even if it is complicated when you look under the hood). The key difference is that the agentic model doesn’t act alone. Instead, it interacts with other specialised AI ‘agents’. Ask a question, and the model directs its ‘agents’ (each a sophisticated model with its own particular focus) to gather and analyse a range of information, which it then assesses, interrogates, clarifies and converts into a meaningful answer.
The model doesn’t ‘think’, as such, but its speed, sophisticated responses, and ability to improve through training are showing incredible potential. Unlike earlier forms of automation, agentic AI tools are designed to be more proactive and autonomous: pursuing defined goals, taking initiative, and adapting to changing contexts.
What could ‘agentic AI’ look like in quantity surveying?
New advances in technology are always exciting, but they should never be an end in themselves. For quantity surveyors, the point of using technology is to transform data into information, knowledge and insight so that we can provide reliable advice about costs, risks, timelines, materials, and the many other parameters of successful projects and portfolios.
Our industry generates huge volumes of data – but data is only valuable if it is accurate, relevant and current, can be rapidly retrieved, and turned into useful information. What agentic AI brings to this is acceleration. It’s a way to speed up the process of accessing and collating data, transforming it into meaningful information, and getting to a point of knowledge.
With AI, what might have taken many hours or days can be done in a matter of seconds or minutes. But is a faster answer a better answer? While speed isn’t everything, it does allow the QS to ask more questions, investigate more angles, and provide a deeper level of insight that enables wiser decision-making.
Take the example of trying to determine how much a 20,000 m2 building would cost. At the most basic level, a QS could analyse a number of projects of that size and determine the averages. With AI, however, we can quickly and easily ask more questions and explore a range of scenarios: What if the building was constructed in 3 years’ time or 10 years’ time? What if it was built in Sydney or Perth or Hobart? We can ask about contingencies, break down the cost of each trade, apply different sustainability targets, understand the depreciation … and the model itself can ‘postulate’ based on its analysis of similar projects, asking further questions or making suggestions in line with the project’s aims.
With this powerful assistant, the QS can access every corner of the available data and provide clients with faster, richer, more complete and more certain answers about risks, feasibility, costs, scheduling and contract administration. In short, by embracing advanced AI, our profession can deliver not only more efficient services, but more strategic value.
The human in the room
As with any major technological shift, agentic AI raises issues of reliability, transparency and accountability in an environment where errors carry real financial and reputational costs. How can a QS – or a client – be confident in trusting AI-generated information or recommendations?
The key point is that AI isn’t replacing professional judgment – it’s informing and reinforcing it. And the human is still in the driver’s seat. Models need to be trained in an ongoing loop of verification, validation, instruction and improvement – guided by the hard-won insight, judgment and perspective of the human construction professional. The human trainer needs to assess whether the model’s answers are accurate and appropriate, explain when and why answers are wrong, and reinforce how to take the right path to get more accurate, coherent and complete outputs.
The strength of this human feedback loop, coupled with the quality of the data source and digital governance, is the root of trust and confidence in adopting advanced AI solutions.
An agentic AI model can help a QS increase their speed, accuracy, rigour and consistency, increasing the visibility of what’s driving costs, where risks sit, and what trade-offs exist. The final judgment, however, still rests with experienced human professionals who understand the project, the market and their client’s commercial goals and have the insights and wisdom that come from years of practice.
In a high-stakes environment where project margins and deadlines are always tight and expectations are always high, every decision matters. Our profession should view AI as not just a shortcut, but a strategic investment in the pathway from data to wisdom. By leveraging the power of advanced AI, QSs can free ourselves to see further beyond the spreadsheet, to think bigger, and to keep providing our clients the smartest advice.
About the author
Ian Baddock is WT Australia’s Chief Information Officer. In this role, he leads WT’s digital transformation and is pivotal in steering the organisation towards the innovative use of data in the construction industry. Ian’s journey began as a quantity surveyor and his expertise in identifying market gaps and crafting data visualisation solutions has been instrumental in adding significant value for clients. Bringing nearly 30 years of diverse experience in the construction sector, his project experience spans health, commercial, retail, aged care, industrial, infrastructure, and education. Over the last two decades, Ian has focused on innovating in data analytics and software development specifically tailored for the construction industry.
