Cleaning the outside windows of a tall building is one of those necessary jobs that is genuinely dangerous and stubbornly expensive. Done the traditional way, it means people working at great height - suspended on ropes, riding platforms and cradles down the face of the building, exposed to wind and weather, doing careful work in a place where a mistake carries the gravest consequences. It is slow, it requires specialist skill and equipment precisely because of the danger, and the risk to the people doing it never goes away no matter how experienced they are. This combination - repetitive work, real physical hazard, high cost driven by that hazard - is exactly where taking the person out of the dangerous position has obvious appeal, and machines that climb a building's glass and clean it have emerged as a practical answer for the routine parts of the task.
A window-cleaning robot works its way across the exterior glass of a building, cleaning as it goes, doing the repetitive scrubbing of large expanses of window without a human being suspended over the edge to do it. The task suits a machine in the way the most convincing practical robotics always does: it is well-defined and repetitive - cover the glass, clean it, move on - and the strongest reason to hand it to a machine is not that a machine cleans better but that the job is dangerous for a person, and a machine at height is not a life at risk. For the large, uniform stretches of glass that make up most of a modern building's face, a machine can work steadily through what would otherwise be long, hazardous hours of human labour aloft.
The value is concentrated in the danger removed and the tedium absorbed, and it is honest about its limits. These machines are well suited to the broad, regular areas of a building; the awkward parts - unusual shapes, obstructed sections, intricate detailing, the spots a machine cannot reach or a stubborn mark it cannot shift - still call for human skill and judgement. The sensible pattern is therefore not to imagine the building cleans itself, but to let machines handle the large, repetitive, dangerous expanses while skilled people deal with the tricky exceptions, ideally from safer positions and for less time in harm's way. The point is not to eliminate the cleaners but to stop asking a person to spend hours suspended over a drop doing work a machine can do just as well from the same place.
As ever, the appeal is narrow and grounded. A window-cleaning robot does one thing - clean the broad glass of a tall building - and its usefulness comes from how dangerous, repetitive and costly it is to have people do that particular part by hand. It does not replace the judgement, the handling of exceptions, or the finishing work that people still do better; it takes the hazardous monotony of the big open expanses. For the owners and keepers of tall buildings, that is a practical and well-bounded use of robotics: safer, because fewer people spend fewer hours at height; steadier, because a machine works through the routine without tiring; and honest, because it leaves to people the parts that genuinely still need them rather than pretending the whole job has been automated away.