A skylight cleaning robot removes fall risk and runs in weather that grounds rope crews, but it cleans at 300-600 m2/h only once set up, and setup alone can take 30-45 minutes. Manual crews win on complex, cluttered or small glass areas.
This is not a technology-versus-people argument. It is a question of where each method is actually better, because both get oversold.
Safety: how big is the real gap?
Falls remain the leading cause of death in construction, and glass roofs concentrate the risk because workers are standing on the surface they are cleaning. A robot removes the person from the roof entirely. That is not a marginal gain. One serious fall costs far more than any machine, in both money and human terms.
The catch: robots bring their own hazards. A 4,000 Pa machine on wet glass can lose its seal at a mullion joint and drop. This is why a certified safety tether is not optional on pitched roofs. If a supplier tells you no tether is needed, end the conversation.
Cost per square metre: which is cheaper?
| Metric | Rope access crew | Skylight cleaning robot |
|---|---|---|
| Cost per m2 | USD 0.8-1.5 | USD 0.15-0.35 after capex |
| Setup time | 60-90 min | 30-45 min |
| Cleaning speed | 80-150 m2/h | 300-600 m2/h |
| Weather window | Wind <10 m/s, dry | Wind <15 m/s, light rain ok |
The robot number only holds if you own the machine and use it often. Rented by the day, a skylight cleaning robot costs about the same as a crew, minus the risk.
Where do manual crews still win?
- Roofs with heavy structural clutter, vents and cable runs that block a robot path.
- Glass areas under 300 m2 where setup dominates the job.
- Detailed frame and gasket cleaning, which needs a hand and a cloth.
- Emergency spot cleaning after a bird strike or construction dust.
A realistic operation uses both. Robots handle the routine full-surface pass every few weeks, and a small crew comes in twice a year for edges, frames and gaskets.
What about water and streaks?
Manual crews can chase a streak with a cloth immediately. Robots move on and leave it, and on frameless glass a missed strip is very visible from below. Running the robot in the early morning reduces this because the glass is cool and water evaporates slower. Lingfeng S1 with its onboard filtered tank is aimed at exactly this routine, lower-spot indoor work.
If your building has hard water above about 200 ppm, robots need filtered supply or the spots get worse than the dirt. Discuss your water source with contact the Lingdu Intelligence team before you buy.
What does a blended programme look like in practice?
A practical programme for a 3,000 m2 glass roof might run a robot every three weeks for the main surface, with a rope crew twice a year for frames, gaskets and the edges the robot cannot reach. The robot does the repetitive, dangerous bulk work. The crew does the detailed, irregular work that needs a human eye.
This split also keeps the crew cost predictable. Instead of six full rope visits costing USD 7,200, you pay for two focused visits at roughly USD 2,400 plus the machine running costs. The saving compounds every year after the machine pays for itself.
Frequently asked question: does a robot clean as well as a person?
On a flat, open glass panel, a robot with a filtered water feed and a clean brush leaves a surface that is hard to distinguish from a hand-finished one. Where it falls short is judgement. A person notices a stubborn patch, a damaged seal or a bird strike and deals with it in seconds.
The honest answer is that robots clean uniformly and humans clean selectively. Uniform is usually what a facade needs, because most complaints come from streaks and missed strips rather than from one dirty corner. If your glass has chronic localised soiling, plan for a human touch-up pass.
What about glass type and damage risk?
Tempered and laminated glass used on roofs is generally hard enough for a robot brush, but coated low-emissivity glass needs care. Abrasive grit trapped in a brush can scratch a soft coating, and on high-performance glazing that scratch is visible and permanent.
Mitigate it by rinsing brushes before each shift and avoiding dry passes on dusty glass. If your glazing carries a specialist coating, confirm with the glass manufacturer that mechanical cleaning is permitted. This is one of the few cases where the cleaning method needs sign-off from the glazing supplier, not just the facility team.
Weigh those constraints honestly before you commit. Our contact the Lingdu Intelligence team can talk you through brush types suited to coated glass.
Does insurance treat the two methods differently?
Increasingly, yes. Insurers ask about working-at-height exposure directly, and a programme that runs a robot for the routine cycle and limits rope access to occasional detailed work presents a lower risk profile than a monthly crew on the roof. Some clients now write this into contracts as a condition of the cleaning scope.
Premiums do not drop automatically just because a machine is on site, but the conversation changes from “how do you stop a fall” to “how do you keep the machine on the glass”. Answer the second question with a certified tether and a trained operator and the underwriting becomes straightforward.
Key Takeaways
- Robots cut fall risk outright, which is the strongest reason to adopt them.
- Expect 300-600 m2/h from a robot versus 80-150 m2/h from a rope crew.
- Manual crews stay better for cluttered, small or detailed glass cleaning.
- Never operate a robot on a pitched roof without a certified safety tether.