Why furnace and kiln relining stays with people
Will AI replace refractory materials repairers? The hero above gives the answer, and the reason sits in the work itself. A furnace, ladle, kiln or oven has to cool down. Then someone climbs inside, chips out burned brick and spent castable, cuts and fits replacement shapes, and places mortar or gunning mix against a curved, cramped wall. Software can schedule that outage and model the heat loss. It cannot get into the vessel.
The second half of the job is judgment in a bad environment. Repairers read wear patterns, decide which courses of lining have to come out and which can wait until the next shutdown, and work in confined space with hot dust and tight permits. Two people can look at the same spalled wall and disagree. The call depends on how the unit is run, what it melts or fires, and how long the plant can afford to be down.
Demand pressure here comes from industry, not from software. This is a small trade: about 1,080 people were employed in it in the United States, with median pay of $61,290 (BLS, 2025). BLS also projects employment falling about 13.7% between 2025 and 2035, which tracks the number of furnaces and kilns running rather than any new tool. Plants in manufacturing keep relining the vessels they have.
What AI does, helps with, and leaves to people
Start with what AI handles on its own. No task in this job’s list sits in the AI-does group, so that share of task time comes to 0%. Nothing in the repair cycle — demolition of old lining, laying brick, patching a tap hole — runs end to end without a person in the vessel.
The assist group is empty too, at 0% of task time. That does not mean the trade uses no software. Thermal imaging, wear logs and maintenance planning systems all sit around the job. They are not yet rated here as taking a share of the named tasks, such as measuring and cutting refractory shapes or mixing and placing castable.
That leaves the person-only group, which our split puts at 100% of task time. Chipping out and removing damaged lining, and rebuilding it to spec inside the unit, both land there. Our coverage figure, which answers whether AI can do the work today, reads 2 out of 100; the coverage method explains how that share of task time is built.
How strong the evidence is
Direct testing is the weak point. The parity question asks whether a machine does this work better than a qualified person, and the evidence grade for that question is D. On this job it means no one has published a head-to-head test of a machine against a working repairer, so we give no parity number at all. Guessing one would be worse than leaving it blank.
What would settle it is specific: a timed, documented trial of robotic demolition and gunning inside a real vessel, on a real outage, with lining life measured afterward against a hand-laid section. Until a plant or equipment maker publishes that, the honest read is untested rather than proven either way. The quality parity method sets out how the grades work and why D carries no score.
Good to know: every figure on this page comes from open data and a published method, not from a survey of people in the trade.
When the picture could change
Most likely after 2045 (8 in 10 of our scenarios). The replacement-year method explains what that window measures and how the scenarios are drawn.
Two things could pull it earlier. The first is machinery built for the vessel rather than the worker: fixed gunning and spraying rigs, and remote demolition arms designed for one furnace shape, which already exist in heavy industry. The second is plant economics. If hourly crew cost keeps rising against equipment cost, a mill with many identical vessels has reason to buy a rig and keep a smaller crew for the odd shapes.
Two things hold it back. The physical share of this work is total, and fixed automation only pays off where the same geometry repeats. Most refractory jobs are one-offs: a different vessel, a different wear pattern, a different access route. The other brake is the setting itself — confined space entry, hot work permits, and the risk of a lining failure in service. Plants sign off on a reline because a trained person vouched for it.
How to stay needed in refractory repair
Lean into the tasks that sit with people. Three are worth building a reputation on: diagnosing wear and deciding what comes out this shutdown, laying and anchoring lining in tight or odd-shaped vessels, and emergency patching that gets a unit back into production without a full reline.
Two skills carry the most weight next to those tasks. One is materials knowledge: which brick, castable or plastic refractory suits the service temperature, the chemistry and the cycle. The other is outage planning — sequencing demolition, installation and dry-out so the plant loses the fewest hours. Both are the parts a plant manager will not hand to anyone unproven.
If you want to look sideways, the nearest work to this trade is Industrial Machinery Mechanics, Maintenance and Repair Workers, General and Brickmasons and Blockmasons. The wider other installation, maintenance and repair occupations family shows how this job sits against its neighbors.
What to check next
This job’s headline figure is 87 out of 100 (higher is safer), and the scoring rules behind it are set out in the methodology. From here, two steps are useful: put this trade against another one you are weighing on the compare tool, or read the list of jobs that mostly need a person to see which other hands-on work lands near it. If you are deciding between trades, the guide to AI and trades careers covers how task erosion shows up in skilled work.