SAFE
Mixing epoxy and cementitious matrices to spec, setting divider strips to a layout line, pouring, seeding aggregate, then grinding, grouting and polishing on your knees in an active job site is work no current robot performs at cost. The only AI-touchable slice is the office layer: bidding, material takeoffs, scheduling and shop drawings for decorative patterns. Real threats to this trade are construction demand cycles, prefabricated terrazzo panels and a shrinking apprentice pipeline — not language models.
Most of this decline happened after 2021 — it is not the pandemic dip.
Median pay $52,180 → $76,170 +16.8% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
-11.1%
Percentage only. The projection counts a different population from the 1,180 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.
Hard to automate, but shrinking anyway
The work resists current AI, yet the BLS projects -11.1% by 2034. Whatever is shrinking this occupation, the evidence does not point to automation — demand, demographics, offshoring and industry decline all shrink jobs that no machine could do.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~100 openings a year on average, including replacing people who leave.
RollerGrinderFinisherInstallerCoat AgentBuildmasterBuild MasterTerrazzo LayerRoofing LaborerTemplate WorkerTerrazzo RollerTerrazzo SetterTerrazzo WorkerTerrazzo GrinderTerrazzo LaborerTerrazzo FinisherTerrazzo MechanicTerrazzo PolisherTerrazzo InstallerConcrete JourneymanTerrazzo JourneymanTerrazzo Tile SetterRustic Terrazzo SetterArtificial Marble Worker
Holding it up: embodiment . Weakest point: liability shield .
Tasks largely resist digitisation Setting brass or zinc divider strips true to a chalk line over an uneven slab, seeding marble chips by hand until coverage reads uniform, then coming back after cure to rough-grind, grout the pinholes and fine-polish — each step is judged by eye and touch on a surface that never repeats, which is why 19 rather than a mixed 12: even the layout you get handed as a shop drawing still has to be adapted to the actual floor.
Hands-on in uncontrolled environments You are on knees and elbows on a green slab, hauling 50-lb bags and epoxy pails, running a 200-lb planetary grinder in silica dust and wet slurry, working around other trades' scaffolds and cords in a building with no HVAC yet — there is no screen-only fraction of this job to discount.
Certification preferred, not legally required No state licenses terrazzo work; you may carry an OSHA 10/30 card, silica-exposure training under 1926.1153 and maybe an NTMA-affiliated apprenticeship certificate, but the bond and the warranty sit with the flooring contractor, and a failed pour is a callback for the company, not a personal license action — hence 5, not 0.
Meaningful discretion You decide whether the slab moisture and crack pattern require a membrane or more control joints, how much resin to catalyze for the shop temperature, and when to stop grinding before you expose too much matrix — irreversible calls on a floor that cannot be lifted, but made inside NTMA and manufacturer specs and with a foreman to back you, so 12 rather than 16.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (19/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (5/20) is whether the law requires a licensed human to sign. Trust premium (12/20) is whether buyers specifically pay for a person. Judgment and accountability (12/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 29 of this occupation's 68 points (43%).
Embodiment (20/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 80/100, still SAFE.
Substrate and moisture calls are already the trade's high-consequence judgment: whether a slab's moisture vapor emission rate and crack pattern warrant a divider strip layout change, a sand cushion versus bonded system, or refusing the pour. If manufacturer warranties on epoxy matrix systems (Terrazzo & Marble Supply, Master Terrazzo) increasingly condition coverage on the installer's documented sign-off of substrate acceptance — as flooring adhesive warranties already do — the installer owns a consequential ambiguous call in writing.
If prefab panel adoption takes the flat, repetitive open-floor pours, the residual field work is disproportionately stairs, cove bases, curved and radius layouts, tie-ins to existing terrazzo and restoration grinding — the tier no template or panel handles. This raises task resistance as a share of remaining work while shrinking headcount, so it is not good news for the occupation's size.
Historic-preservation and restoration work already commands a hand-craft premium: if more state historic tax credit programs and federal Secretary of the Interior standards reviews explicitly require in-situ poured terrazzo restoration by trained craftspeople rather than prefabricated panel replacement, the human-made distinction gets written into the specification, not just the sales pitch. Also visible: NTMA (National Terrazzo & Mosaic Association) contractor certification being named in architect specs, which buyers pay a premium to satisfy.
Terrazzo has no personal license. The realistic route is indirect: epoxy resin matrices are subject to VOC and isocyanate handling rules, and if OSHA's respirable crystalline silica standard (1926.1153) enforcement on grinding and polishing tightens to require a named competent person with documented exposure-control responsibility on each pour-and-grind operation — the way lead and asbestos abatement already require certified supervisors — a person becomes legally required on site and personally exposed to citation.
The limit. Already 68 with task_resistance and embodiment near maximum; the headroom is only in the three institutional dimensions and totals maybe 10-12 points. The binding constraint is not AI at all — it is 1,180 workers, cyclical commercial construction demand, panel prefabrication, and apprentice supply. A higher register score would not change any of those.
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 68. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.