Skip to content
man bot

Elementary occupations · code 9123

Window cleaners

Window cleaners wash and polish windows and other glass fittings. Tasks include - (a) washing windows or other glass surfaces with water or various solutions, and drying and polishing them; (b) using ladders, swinging scaffolds, bosun’s chairs, hydraulic bucket trucks and other equipment to reach and clean windows in multistorey buildings; (c) selecting appropriate cleaning or polishing implements. Examples of the occupations classified here: - Window cleaner

None of this job's tasks scored high enough to count. It ranks 21 of 427 for holding up against AI.

Also known as

  • window cleaner

Job titles from the EU's ESCO list. Search for any of them on the jobs page and you'll end up here.

88.9
Holds up
against AI Not Exposed

AI could take on 11.1 of 100, on average
so it holds up at 100 − 11.1 = 88.9

That's the whole sum, and you can check it against the ILO's study.

53
How hands-on
out of 100

From O*NET, matched to one US job

31.7
How fiddly
and cramped

finger skill 35.7 · hand skill 34 · cramped spaces 25.5 · averaged

Kind of hands-on work · hands-on 50+ · fiddly under 39.4 Hands-on, less fiddly work

Less fine handwork and fewer cramped spaces than a typical hands-on job. That clears one hurdle for machines. It doesn't mean machines are doing this work yet.

estimated O*NET 31.0 BLS occupational crosswalk chain ISCO-08 to 2010 SOC (Aug 2012, rev. Jun 2015); 2010 to 2018 SOC (Nov 2017); O*NET-SOC 2019 taxonomy data 0.7-work-setting measures chosen by Frey & Osborne (2017) · the grouping is a sorting, not a measurement

This number is probably too high

This score only looks at AI tools like chatbots. It doesn't look at robots at all — and this job rates 53 out of 100 for how hands-on it is, which is exactly where robots come in.

Have a look at the tasks below and judge for yourself. The ILO scored them low because a chatbot can't do them. Whether a machine with hands could is a different question, and we can't answer it yet.

The tasks behind the score

The score is the average of these. A job is really a bundle of tasks, and an average can hide how uneven that bundle is.

The task AI could do most of scores 15, while the average is 11.1. A big gap means some of this job is wide open to AI and some of it barely at all. That's quite different from a job where everything sits somewhere in the middle, even when the averages match.

  1. 15

    Selecting appropriate cleaning or polishing implement.

    Very Low
  2. 9.5

    Washing windows or other glass surfaces with water or various solutions, drying and polishing them

    Very Low
  3. 8.9

    Using ladders, swinging scaffolds, bosun´s chairs, hydraulic bucket trucks and other equipment to reach and clean windows in multistorey buildings

    Very Low

3 tasks · the High to Very Low labels are the ILO's

from the source Working Paper 140, Generative AI and Jobs 2025 index data 0.7-work-setting

Helped, or replaced?

The score above can't tell a tool that makes you faster from one that does the task instead of you. Jobs and Skills Australia looked at those two things separately, and this is their result, matched to this job.

One Australian job matches this one closely, so this is about as reliable as it gets here.

AI helps you
44
AI does it for you
15

Helping wins by 29 points.

estimated Our Gen AI Transition (Gen AI Capacity Study) September 2025 detailed data release data 0.7-work-setting matched to Australian jobs 8116

Compared with similar jobs

There are 33 jobs in the “Elementary occupations” group, averaging 84.9. This one is above that.