Share

Australia leads the world on construction AI adoption. Productivity confidence has not followed.

Australian construction reports artificial intelligence adopted at scale at nearly twice the global rate. It is also less confident than the rest of the world about improving productivity. The distance between those two numbers is the whole story. Australian construction businesses are adopting artificial intelligence faster than their global peers. That is not the finding […]

Read

Mon 28 Sep 26 6:00:00 AM

tgb-logo-crop

Australian construction reports artificial intelligence adopted at scale at nearly twice the global rate. It is also less confident than the rest of the world about improving productivity. The distance between those two numbers is the whole story.

Australian construction businesses are adopting artificial intelligence faster than their global peers. That is not the finding most people expect, and it is the first thing that makes the productivity conversation harder than it looks.

In KPMG’s Global Construction Survey 2025/2026, 43.8 per cent of Australian respondents reported AI adopted at scale. The global figure was 24 per cent. Australia is not at the back of this queue. It is a long way toward the front.

Then comes the number that sits awkwardly beside it. Australian leaders in the same survey were less positive about improving productivity than their global counterparts, at 68.8 per cent against 75 per cent. On revenue growth, attracting talent and managing risk, Australia matched or beat the global result. Productivity was the one place it fell behind.

So this is not an industry refusing to adopt technology. It is an industry adopting technology at pace and remaining unconvinced about the thing technology is supposed to deliver.

The gap is between ambition and execution, not between builders and software

KPMG Australia attaches a figure to that gap. Its deal advisory partner Amanda Coneyworth describes a potential annual opportunity of A$56 billion if construction productivity matched the economy wide average, and identifies fragmented delivery, siloed data and limited technology adoption, particularly among mid tier contractors, as what sits in the way.

The figure is an estimate of distance rather than a cheque waiting to be collected. It measures how far construction sits from the rest of the economy. Nothing about it is automatically recoverable by buying another platform.

Investment in digital tools is already widespread across the industry

Infrastructure Australia’s 2025 Infrastructure Market Capacity Report found that 64 per cent of building and construction firms surveyed had invested in digitalisation in the past year, with that spending spread across building information modelling, artificial intelligence, robotics and automation.

The same report recorded multifactor productivity rising 2.0 per cent in 2023 to 2024, reversing the decline of the year before, while noting that the long term trend remains flat and still sits well below where it was in the mid 1990s.

Both statements hold at once. The money is going in. The long run line has barely moved.

Multifactor productivity is not the same as labour productivity

Labour productivity measures output against hours worked. Multifactor productivity measures output against labour and capital combined. A business can lift output per hour by adding equipment, software or systems while multifactor productivity stays flat, because the money spent on those things counts as an input too. This is why an industry can invest heavily, report better output per hour, and still show a flat long term productivity trend.

Digitising a process is not the same as changing it

A business can buy a platform without altering the process behind it. A paper form becomes a digital form. A spreadsheet becomes an app. Email threads move into a project management tool. An estimator upgrades to more capable software and prices the same way.

The technology is newer. The process is not. What results is a digital version of the same inefficiency, and it can be harder to spot than the paper version was, because it looks modern and it produces reports.

That is the reading KPMG’s own numbers invite. High adoption reported alongside persistent execution problems suggests the constraint is not access to tools.

Coordination is where construction time actually goes

A single residential project can carry a builder, a client, a designer, an engineer, a certifier, a surveyor, suppliers and a long list of subcontractors. Every design change, product substitution, client selection, variation and inspection creates information that has to reach the right people among them.

Infrastructure Australia treats subcontracting as a structural feature of the market rather than a detail, putting it at 41 per cent of infrastructure construction and noting the industry’s own view that it brings interface risks on large projects, higher supervision costs and reduced training opportunities.

Residential work runs a smaller version of the same arithmetic. The more interfaces a project carries, the more chances information has to be lost, duplicated or misread. Software moves that information faster. It does not reduce the number of handovers.

Siloed data is a different problem from missing data

Siloed data sits in KPMG’s list of what holds Australian construction productivity back, and it is worth separating from a shortage of data. Construction businesses generate an enormous amount of information: costs, labour hours, subcontractor performance, materials, schedules, variations, defects, procurement and margins.

Software moves that information faster. It does not reduce the number of handovers.

Knowing what a project cost is not the same as knowing why it cost that. Knowing a job ran three weeks late is not the same as knowing which recurring step caused the delay. Knowing one subcontractor produced more defects than another is not the same as knowing whether the cause was workmanship, design, sequencing or a communication breakdown upstream.

When those records sit in systems that do not talk to each other, the same detail gets entered more than once and the copies drift apart. The site works from one version. The office finds the other one later.

Part of the productivity problem sits above the individual builder

Infrastructure Australia points at procurement itself. It describes current models, focused on lowest cost and project by project tendering, as discouraging the adoption of scalable innovations including modern methods of construction, digital engineering and low emissions materials.

The effect shows up in the numbers. Prefabricated construction still makes up less than 5 per cent of the total market, and Infrastructure Australia notes that regulatory and financing barriers may also be slowing its uptake.

That matters because no single business controls the conditions it works under. Where every project is treated as a one off, where information standards change between clients, and where risk is pushed down the chain, improvement at the company level gets harder no matter what software is installed.

The pipeline is not easing either. Infrastructure Australia’s Annual Budget Statement 2026 puts the major public infrastructure pipeline at $242 billion over five years, up 14 per cent, against a peak workforce shortage that could reach 300,000 workers by 2027. Productivity stops being an abstract debate when the same labour pool is being asked to deliver more.

What the A$56 billion figure actually describes

Underneath the headline number sit distinctly unglamorous items. Less rework. Less waiting. Fewer duplicated entries. Better sequencing. More predictable procurement. More use of the data a business already holds. A drawing reaching the right person before the work is scheduled rather than after.

Some of those involve technology. Many do not. That is the uncomfortable half of the KPMG finding, and it is also the encouraging half, because it means the constraint is not capital expenditure.

Which points at a different question from the one most technology conversations start with. Not which software a business needs, but where the business is losing time, and whether anything on the market actually removes that loss. On the evidence in both reports, Australian construction has already answered the first question at scale. The second one is still open.

The Good Builder Take

Australia is not behind on adoption, and that is the part of this data worth sitting with, because most of the commentary assumes the opposite. The builders in our community who report real gains from technology tend to describe the same order of events. They found the step that was costing them time, then went looking for something that removed it. The ones who went the other way, buying the platform first and hunting for a use afterwards, generally ended up with a newer, more expensive version of the same problem. Neither report says technology does not work. Both suggest the sequence matters more than the spend.

Frequently asked questions

What is the A$56 billion construction productivity figure?

It is KPMG Australia’s estimate of the potential annual opportunity if Australian construction productivity matched the economy wide average. It was published alongside the Global Construction Survey 2025/2026 as commentary from KPMG Australia rather than as a survey result, and it describes a gap rather than recoverable revenue.

How does Australian construction AI adoption compare with the rest of the world?

KPMG’s Global Construction Survey 2025/2026 recorded 43.8 per cent of Australian respondents reporting AI adopted at scale, against 24 per cent globally. Australia also reported a higher share of transformation spending going to people, at 23.6 per cent against 21 per cent globally.

Is construction productivity in Australia improving?

Infrastructure Australia recorded multifactor productivity rising 2.0 per cent in 2023 to 2024, which reversed the decline of the previous year. It also reports that the long term trend remains flat and sits well below mid 1990s levels, so a single year of improvement does not yet change the direction of travel.

Why does technology investment not always improve productivity?

KPMG identifies fragmented delivery, siloed data and a gap between ambition and execution sitting alongside high adoption rates. Where a digital tool reproduces an existing process rather than changing it, or where separate systems do not share data, the underlying inefficiency stays where it was and simply moves into a newer interface.

What is slowing modern methods of construction in Australia?

Infrastructure Australia points to procurement models focused on lowest cost and project by project tendering, which it says discourage scalable innovations including modern methods of construction and digital engineering. It also notes regulatory and financing barriers. Prefabricated construction remains under 5 per cent of the total market.


Related articles

Last updated 22 September 2026. Figures drawn from KPMG’s Global Construction Survey 2025/2026 and Infrastructure Australia’s 2025 Infrastructure Market Capacity Report, published 13 November 2025, and Annual Budget Statement 2026, published 1 April 2026.

General information only. This article reports publicly available research and government data and does not take account of the circumstances of any particular business. It is not financial, legal or professional advice.


TGB Editorial
Author: TGB Editorial

0 Comments

TGB Editorial

TGB Editorial

Related News

Australia Is Repairing Almost as Many Homes as It Is Building

Australia Is Repairing Almost as Many Homes as It Is Building

Declared events damaged 145,392 homes in 2025. The country completed 172,657 new ones. The repair market draws on the same trades, the same materials and the same capacity, and the official housing numbers do not separate the two. In 2025, declared natural disaster...

TRENDING

BROWSE FURTHER