A machine that does a job four times faster than a crew does not necessarily finish the building four weeks earlier. It only shortens the programme if the job it replaced was holding everything else up. That single distinction explains why two contractors can buy the same equipment and report completely different results.
Four things to hold on to
- Ask which activity currently drives the completion date before evaluating any machine.
- Robotic layout covers roughly four to five times the area per shift that a manual crew does.
- Automated rebar tying has compressed twelve-day mat sequences to four to six days on documented deployments.
- The largest recoverable losses are not slow tasks. They are errors that force work to be done twice.
Speed only counts where the critical path is
The critical path is the chain of activities that sets the completion date. Shorten a task on it and the building finishes sooner. Shorten a task with float, and you have simply created a crew waiting for the next trade. This is not a subtlety for planners; it is the entire economics of a robot purchase.
The practical test is unglamorous. Pull the last three programmes, find what was actually driving the date at each stage, and ask whether any of the automated workflows on the market touch those activities. On a concrete-framed building the answer is often yes, because the reinforcement and layout sequences sit directly on the path. On a fit-out heavy scheme the answer is frequently no, and the money is better spent elsewhere.
The four workflows where machines are genuinely running
The Construction Robotics Report published on 5 March 2026 by Zacua Ventures with Hilti Ventures and 94 Ventures is the most useful recent stocktake, because it separates repeatable production from demonstration. It identifies four workflows with material deployments: layout and measurement, groundworks and earthmoving, structural and rebar work, and inspection with digital capture.
| Workflow | Reported performance | Effect on the programme |
|---|---|---|
| Layout and measurement | 40,000 to 70,000 sq ft per shift, against 8,000 to 15,000 for a manual crew | Direct, and it unblocks every following trade |
| Structural and rebar | 300 to 450 tie intersections per hour, against 40 to 80 by hand | Direct on concrete frames, where the mat gates the pour |
| Groundworks and earthmoving | Autonomy kits for trenching, grading and piling | Variable, and highly dependent on ground conditions |
| Inspection and digital capture | $3,500 to $5,500 per month on a $50 million project | Indirect. It protects the programme rather than shortening it |
The rebar figure is the one that translates most cleanly into days. On a mat of around 200,000 intersections, the report describes a sequence that would run twelve days by hand completing in four to six days, with the pour and everything after it pulled forward by the same margin. That is a schedule gain, not a productivity statistic, and the two are constantly confused in sales material.
There is a caveat in the same research worth quoting to anyone planning a first deployment: skipping the pre-deployment geometry review, which sorts the regular zones from the awkward ones, is the most common reason a robotic rebar programme misses its modelled schedule. The machine performs. The plan built around it does not.
The days that get lost, not spent
Most programmes do not overrun because tasks were performed slowly. They overrun because a proportion of the work had to be done twice. The Get It Right Initiative, the UK industry research body set up to quantify exactly this, has put the cost of avoidable error at somewhere between 10% and 25% of project cost once indirect consequences and latent defects are counted, with the directly attributable share nearer 5%. A 2022 study by Autodesk with FMI reached a comparable conclusion from the other side of the Atlantic, estimating rework at around 5% of United States construction spending, which on a $1.3 trillion base works out above $65 billion a year.
This is where digital capture earns its keep. Consistent weekly scanning does not lay a single brick faster. What it does is fix what was in place on a given date, which turns a six-week argument about who moved a duct into a five-minute look at the record. The March 2026 report puts the avoided-claim value on a $50 million project at $80,000 to $150,000 against a monthly cost in the low thousands, with payback inside six months. Those are the days nobody plans to lose and everyone does.
What automation does not compress
Being clear about the ceiling matters as much as the gains. Several parts of a building programme are indifferent to how automated the site is:
- Statutory approvals and permitting. These run on institutional time, and no machine on site influences them.
- Material and plant lead times. A twenty-week switchgear order is twenty weeks regardless of how quickly the risers were laid out.
- Curing, drying and settlement. Chemistry does not accelerate to suit a Gantt chart.
- Commissioning and handover. Often the most compressed phase on a job, and the least automated.
Where genuine multi-week compression does appear, it usually comes from moving work off the site entirely into a factory rather than from any single robot on the plot. That is a different decision, with different risks, and it has to be taken at design stage rather than at procurement.
How to evaluate a machine without the brochure
Three questions cut through most vendor conversations. First, which activity on our current programme does this touch, and does that activity have float? Second, what is the mobilisation time, and how many square metres or intersections does the site need to offer before the machine is cheaper than the crew? Third, what happens on the days it does not work, and who absorbs that.
The safety picture deserves its own assessment rather than a line in the same business case. Automation removes people from some hazards and introduces others, and the applicable duties remain those of the health and safety regime in the jurisdiction of the site, backed by proper training, never a supplier’s method statement alone. We have looked at that trade-off separately in our piece on how automation changes construction site safety.
Does automation actually reduce the number of workers needed?
On the specific task, yes, and the crew is usually redeployed rather than removed. On the documented rebar deployments the labour saving is real but partial, because the machine still needs an operator, a geometry review and people finishing the zones it cannot reach.
What size of project justifies robotic layout?
The 2026 research puts viable economics above roughly 150,000 applicable square feet per month. Below that, mobilisation and subscription costs outweigh the productivity gap, and a competent crew with a robotic total station remains the better answer.
Is it faster to automate an existing process or redesign it first?
Redesign first, in almost every case. Automating a sequence that was arranged around manual constraints tends to lock those constraints in place. The teams reporting the largest schedule gains changed the order of work to suit the machine, rather than dropping the machine into an unchanged programme.
The wider shift behind the machines
Site robotics is one strand of a broader move toward automated design, construction and building operation.
