Every estimate, variation, delay and defect a building business records says something about how it actually performs. The widening gap is between the builders who read that record and the builders who file it.

A building business generates information constantly. Every estimate says something about pricing. Every variation says something about scope, communication and the decisions a client made under pressure. Every finished home leaves a record of build time, labour, procurement, defects, rework and whatever margin survived to the end.

Most of it is treated as paperwork. It gets recorded, filed, archived when the job closes, and never looked at again as a set.

That is the part that is shifting. Not because builders want to run technology companies, but because the information already sitting inside a building business is starting to be worth more than the software bought to collect it.

Every project a builder finishes already produces a data trail

Estimating and tendering. Lead sources and sales conversion. Project pricing and gross margin. Variations. Construction time. Labour utilisation. Subcontractor performance. Procurement and material costs. Delays, defects, rework, warranty claims, cancellations, cash flow, supplier reliability, house designs and customer type.

Nearly all of that already exists somewhere inside a typical building company. So the question is not whether the data exists. It is what happens to it once the job is closed.

A variation can be a variation. Or a business can know which types of variation recur, why they happen, and what each one costs in margin.

A delay can be a delay. Or a business can see the causes that keep repeating across a hundred jobs.

A subcontractor can finish a job. Or a business can hold a record of that subbie on time, cost and quality going back years, which is a different basis for how builders manage subcontractors than memory and goodwill.

The productivity numbers are why this is being discussed now

Construction employs around 1.3 million people, on ABS figures for the 2023 to 2024 financial year. It is also one of the industries where output per unit of input has been going backwards.

The 2025 Infrastructure Market Capacity Report recorded construction multifactor productivity rising 2.0 per cent in 2023 to 2024, while describing the longer term trend as flat and below mid 1990s levels. The ABS has since published the following year. In the 2024 to 2025 financial year, construction multifactor productivity fell 2.8 per cent and labour productivity fell 2.9 per cent.

Productivity is not a measure of how fast anyone swings a hammer. It measures how much useful output an industry gets from the people, materials, capital and time it puts in. Information is part of that conversion.

If a business loses the same fortnight every year to the same design issue or the same interface between two trades, the record needed to see it is usually already there. What is missing is anyone looking at it together.

Multifactor productivity

Multifactor productivity measures output against the combined labour and capital an industry uses to produce it. It is not a measure of individual effort. A fall means more labour and capital went in for roughly the same value out. ABS reported construction multifactor productivity fell 2.8 per cent in the 2024 to 2025 financial year, driven by flat gross value added alongside a 3.0 per cent rise in hours worked.

Having software is not the same as having intelligence

The industry has spent a decade digitising. Estimating software, project management platforms, CRMs, accounting systems, procurement, scheduling, client portals, safety and quality systems. And now AI.

Infrastructure Australia found 64 per cent of the building and construction firms it surveyed in 2025 had invested in digitalisation in the previous year, across building information modelling, AI, robotics and automation. Another 26 per cent invested nothing, and the report noted small business lagging on both adoption and awareness of AI tools.

But a company can run ten systems and still struggle to answer one question. Why is this type of project making less money than it used to?

That is the distance between holding data and using it, and it is where most of the practical conversation about technology and AI for builders now sits.

Fragmented delivery is what stops the numbers adding up

A single project can involve the builder, the client, an architect, engineers, certifiers, suppliers and a long list of subcontractors. Each produces information, in different systems, recorded differently. Some of it is structured, some sits in spreadsheets, some is buried in email threads, and some exists only in the head of whoever was on site that day.

The KPMG Global Construction Survey 2025/2026 identifies fragmented delivery, siloed data and limited technology adoption among Australian firms, particularly mid tier contractors, as a driver of poor construction productivity. The same research ranks improving data and technology capabilities as the second most cited strategic priority for Australian respondents, behind managing risk.

KPMG also puts a number on the gap. It estimates a potential A$56 billion annual opportunity if Australian construction productivity matched the economy wide average.

That figure is not cash sitting in builders’ accounts waiting to be swept up. It is a modelled estimate of economic value attached to closing a productivity gap, and it is useful mainly as an indication of scale.

A building company can know everything that happened on a thousand projects without knowing what those projects are telling it collectively.

The most useful numbers are usually the ones nobody reports

Building businesses track the headline figures. Revenue, contracts signed, homes under construction, average contract value.

The numbers that explain those figures tend to sit further down. How often a job stopped while it waited on information. How often material had to be reordered. How many variations came from something that should have been settled at design. How many hours went into correcting work. Which designs quietly absorbed management attention.

None of that looks impressive in a capability statement. Collectively it is a map of where a business spends resources it never invoiced for.

Consistency beats sophistication

There is a version of this conversation that ends with every builder needing predictive analytics. The evidence does not support that.

A business does not need a model to work out that the same problem has come up dozens of times. It needs that problem recorded the same way each time.

Which sounds basic, and is where most attempts come apart. If one supervisor logs a hold up as a supplier issue, another logs it as materials late, and a third does not log it at all, there is a data problem underneath the technology problem.

Our reporting on small business productivity in construction found the same pattern. The recent gains have come from tightening process, with digital tools in a supporting role rather than a starring one.

AI moves the advantage to whoever kept the better records

AI does not manufacture business intelligence out of nothing. It needs information that is accessible, structured and reasonably reliable.

So the question shifts. Access to AI will not be the differentiator for long, because access is becoming universal. What differs is what each business can hand it.

Five years of clean project history gives a system something to work with. Five years scattered across inboxes, spreadsheets and three platforms that do not talk to each other gives it very little.

It is worth noting what the builders who went all in on AI and then pulled back reported. Their limits were about scope, about which tasks a tool can actually carry. The quality of what it is given to work with is the other half of the same question.

The businesses with the advantage will not look like data companies

They will look like builders. Project managers, supervisors, estimators, sales, trades and subbies. Same weather, same supply chains, same clients, same site conditions.

The difference sits behind all of that. Knowing where time goes and where margin goes. Which problems recur. Which suppliers hold up. Which designs work operationally rather than on paper.

For most of the industry’s history, information has been project centric. The job finishes, the file closes, the business moves on. Treating those closed files as a continuous record of how a business behaves is a different proposition, and it sits closer to the fundamentals of running a building business than to anything on a software roadmap.

It is also not limited to volume builders. A small operator who understands its own performance in detail is holding something a much larger competitor may not have. And unlike a platform, that understanding cannot be bought with a subscription. It has to be accumulated, which is exactly why it is worth something.

THE GOOD BUILDER TAKE

The industry has spent ten years buying systems. The next stretch is about what those systems have quietly been recording the whole time. The starting point is not a data strategy or a new subscription. It is recording the same problem the same way twice, which is the one part no vendor can do for a building business.

Frequently asked questions

Sources

Australian Bureau of Statistics, Estimates of Industry Multifactor Productivity, 2024-25

Australian Bureau of Statistics, The nuts and bolts of the Australian construction industry

Infrastructure Australia, 2025 Infrastructure Market Capacity Report

KPMG Australia, Global Construction Survey 2025/2026

Last updated 22 September 2026. Productivity figures current to the ABS release of 6 February 2026.

General information only. This article reports on publicly available data and industry research and does not take into account the objectives, financial situation or needs of any particular business. It is not legal, financial or professional advice.

Author: TGB Editorial