Analysis
Operating costs in construction — where AI savings come from and how to measure them
The AI Business Center Foundation (AIBC) is a think tank and advisory partner that helps the boards of construction, industrial and technical-services companies adopt artificial intelligence so that it delivers measurable results and builds competitive advantage. This analysis explains where AI savings in construction come from and how to measure them honestly.
A single savings figure is no longer convincing to boards. Measurement in one process has become the argument
In spring 2025 this text carried a title that promised to cut operating costs in construction by a specific percentage range. We have changed it. The range came from selected examples and market forecasts with no described methodology, and a board that commits to an implementation on the strength of such a figure has no way of holding it to account.
The problem is not confined to construction. According to McKinsey, The State of AI 2026, 8 in 10 employees see time savings from GenAI; only 4 in 10 companies see an impact on operating profit (EBIT). Between the claim and the result lies measurement.
Where the savings come from
Artificial intelligence does not reduce costs by itself. It reduces them where it removes waste, errors and manual work built into the everyday running of a construction site:
- Preparing bids and cost estimates. Less time spent on quantity take-off and costing, and fewer errors that only surface on site.
- Schedule and resources. Earlier signals about phases at risk of delay, and better planning of people and equipment.
- Quality control. Comparing the work as executed with the design and documentation before an error is built in.
- Equipment maintenance. Planning inspections according to the condition of machines, not only the calendar.
- Fewer corrections, disputes and penalties. Every change that reaches the right person in time is a cost that never arises.
In each of these areas the condition is the same: a described process, an owner, and data in a form that can be analysed. Without them a tool speeds up the production of documents but does not reduce costs.
How to measure the effect
- Choose one process of high financial significance that recurs on every site.
- Measure the baseline using data from the company’s systems: time, number of corrections, number of changes priced after the fact.
- Appoint an owner within the company who is responsible for the process and for the measure.
- Run a pilot with a measurement afterwards, and compare the effect with the cost of implementation and maintenance.
- Decide: scale up, improve or stop.
Before a pilot, we state potential only as a range*, never as a promise.
* Values depend on the process; confirmed in the audit.
What has changed since first publication
- We have removed from the title and the text the percentage savings ranges that had no described methodology.
- We have removed examples of implementations at specific companies, together with amounts and percentages we could not confirm in primary sources, and a market-value forecast taken from a press release.
- We have added McKinsey data from 2026 on the gap between reported productivity and the impact on results.
- We have added a way to measure the effect in a pilot.
We develop the measurement method further in the AI value report for the board, which is part of semantic transformation, that is, organising a company’s names, processes and data so that people, language models and board reports understand them the same way.
Frequently asked questions
How much can a construction company save with AI?
There is no honest single figure for the whole industry. The effect depends on the process, the starting point and the quality of the data. What can be relied on is a result measured in a pilot on the company's own data: the situation before, the situation after and the cost of implementation.
Where should a company start so that savings can be measured?
With one process of high financial significance, with an owner and a baseline measure: for example bid preparation, change management or quality control. Then a pilot measured before and after, and a board decision on scaling up.
Let’s start with a conversation
Let’s talk about one process in which it is worth measuring savings.