Why this job stays next to the toolbox
A helper’s day is organized around someone else’s hands. You carry materials to the job, hold a part steady while a mechanic bolts it down, hand over the right tool at the right second, and clean the area so the next step can start. That rhythm is physical, unplanned, and shared with another person. Software does not get a grip on it.
The second reason is the setting. Helpers work in crawl spaces, plant aisles, truck bays, and half-finished rooms. Nothing is fixed in place. A ladder has to be footed on uneven ground. A panel cover goes back on with the gasket seated right. The job changes shape hour by hour, and the person leading the repair expects a second pair of eyes and arms that can adjust without being told twice.
Pay and headcount give the practical frame. About 95,580 people work in this occupation in the US, with median pay of $39,630 a year, and employment is projected to grow 5.3% from 2025 to 2035 (BLS, 2025). That is steady demand for hands on site, much of it in construction and site work and in plants and vehicle shops.
What software handles, what it assists, and what it does not touch
The share of task time AI can take on its own is 0%. What sits there is paperwork at the edge of the job: logging what was done on a work order, and pulling up the right manual, part number, or spec sheet before a repair starts. Those are text-and-lookup tasks, and they are the ones language tools handle cleanly.
The assisted share is 6%. Assistance here means a phone or tablet in your pocket rather than a machine on the floor. A scheduling app routes you to the next site. A parts system flags what is in stock before you load the truck. You still make the trip, open the box, and carry the load.
Everything else, 94% of task time, stays with people. That includes holding and positioning equipment while it is fitted, moving tools and materials between the truck and the work point, cleaning and prepping the area, and disassembling parts so the lead worker can inspect them. The coverage figure of 5 reflects that split; the coverage method explains how task time is weighted.
What has actually been tested
No study has put an AI system against a helper on a real job and measured the result. The evidence grade for this occupation is D, which is our mark for work that has not been directly tested. So there is no parity number on this page, and there should not be one.
What would settle it is specific: a trial where a mobile robot works a full shift alongside a trade crew, carrying and positioning materials in a space it has never mapped, and someone records completion rates, damage, and how often a person has to step in. Benchmarks run in offices or on text tasks say nothing about footing a ladder. Our quality parity method and the wider scoring method set out why an untested job gets a grade instead of a number.
When the picture could shift
Most likely after 2046 (8 in 10 of our scenarios). The replacement-year method explains what that window is measuring and how the spread is built.
Two things could pull it earlier. Hardware is the first: the robotics panel above puts nearly all of this job’s task time in the physical column and tags the capability needed as mobile robots, so steady progress in legged and wheeled machines that handle clutter matters more here than any language model. The second is cost. The monthly cost comparison on this page shows software priced far below a wage, and if general-purpose machines ever land near that line, employers in large plants will test them.
Two things hold it back. Sites are unstructured and change daily, which is the blocker that machine vision and navigation keep tripping over. And a helper’s value is partly social: knowing what the lead tech needs before being asked, and being the person who can be sent for anything. Replacing one slice of that leaves the rest undone.
What to do: treat the helper role as a doorway, not a destination, and start logging hours toward a license or apprenticeship early.
How to stay needed in this role
Lean into the tasks that sit in the human column. First, equipment handling: get good at rigging, lifting, and positioning heavy parts safely, because that is the part of the job nobody is automating on a live site. Second, disassembly and setup: learn how machines come apart in the right order so inspection goes faster. Third, site readiness: materials staged, area clean, tools where the lead can reach them. Crews keep the helper who removes friction.
Two skills raise the ceiling. One is diagnostics, even at a basic level, so you can describe a fault clearly instead of just reporting a noise. The other is comfort with the tablet side of the job, including work-order systems and parts lookup, since the tasks AI already handles are the ones you want off your plate.
Where helpers go next is usually a licensed trade. Close neighbors worth reading are maintenance and repair workers, general, industrial machinery mechanics, and helpers–electricians. The rest of the group sits on the other installation, maintenance, and repair occupations page. You can put any two of them side by side with our job comparison tool, or see how hands-on roles stack up on the jobs that mostly need a person list.