■ HIGH RISK ■ Construction & Trades
Largely, yes, for the open-floor grunt work — autonomous and semi-autonomous sanders are built for exactly that. The edges, stairs, repairs, and finish judgment stay human longer, but they're a fraction of the billable hours.
“Robotic sanders don't leave swirl marks or charge by the hour.”
Our AI replacement risk score — how we score jobs
Sanding a hardwood floor is a sequence: assess the wood and old finish, pick grit progression, run the big drum or belt machine in overlapping passes with the grain, switch to an edger along walls, hand-scrape corners and stair nosings, vacuum obsessively, then screen between finish coats. The skill is in reading the floor — knowing when cupped boards need an angled first cut, when an old wax finish will gum the paper, when one more pass burns through the wear layer of an engineered floor.
The main event, though, is hours of walking a heavy machine across open floor in straight lines. That's a navigation problem, and robotics has solved worse. Autonomous floor-sanding machines already map a room and grind it flat with more consistent pressure than a tired human at 4 p.m., and construction robotics firms have targeted concrete grinding and surface prep for the same reason: it's dusty, repetitive, injury-prone, and geometrically simple. Contractors adopting them keep the machine running while one worker does edges — one crew where two used to be.
What resists automation is everything that isn't open floor: staircases, herringbone borders, board replacements, water-damage repairs, blending a patch into a ninety-year-old oak floor, and the finish-selection conversation with a homeowner. Occupied homes full of furniture also slow the robots down. Our risk score of 74 reflects a trade where the machine takes the square footage and the human keeps the perimeter — and square footage is what pays. The refinishers who prosper will be the ones invoicing for judgment while the robot does the walking.
Automatability: our editorial assessment of current and near-term AI capability
Adoption is underway in commercial surface prep now and moves into residential refinishing as machine costs drop — expect serious crew-size pressure by 2030. The pattern will be shrinkage, not disappearance: one skilled finisher supervising a machine plus doing edges and repairs, where a two- or three-person crew used to bill. Solo operators serving small residential jobs feel it last.
Yes — autonomous grinding and sanding units are in commercial use for concrete and surface prep, and hardwood-capable machines exist. They excel at large open areas and consistent flatness. Residential adoption is slower because homes are cluttered and jobs are small, but rental availability will change the math for contractors this decade.
Safe-ish if you're more than a machine driver. The open-floor sanding hours are the automatable core, and they're most of the invoice. Finishers who can repair, patch, stain, and restore — and who adopt the machines rather than compete with them — will still have work. Pure grunt-sanding labor won't age well.
Judgment and edges. Reading whether an engineered floor has one sanding left in it, blending a repair into aged oak, hand-detailing stairs, and talking a client out of a finish that will show every dog hair — that's human work for the foreseeable future. The straight passes across the living room are not.
If you do regular commercial or large-format work, run the numbers now — labor saved on big open floors pays the machine off fastest. For small residential refinishers, renting when large jobs appear beats owning. Either way, being the operator beats being the competitor.