Why the work stays close to the bench
A wood patternmaker turns a drawing into a physical master. That master becomes the mold used to cast a metal part. The job is layout, cutting, shaping, assembly, and fit. Software can draw the shape. It cannot sand a draft angle until the pattern pulls clean from sand.
Two parts of the work explain most of the answer. First, shaping and finishing stock by hand and machine, where the patternmaker feels grain, reads tear-out, and adjusts as the piece changes. Second, fitting and verifying the assembled pattern, checking dimensions against the specification and correcting what does not sit true. Both are judgment applied to one piece of wood in one shop.
Scale matters too. The Bureau of Labor Statistics counted about 220 wood patternmakers in the United States, with median pay of $49,630 (BLS, 2025). A trade that small rarely attracts purpose-built machines. Nobody writes a robot for 220 benches.
What AI does, what it assists, and what people keep
Start with the computational end. Calculating shrinkage and machining allowances, and laying out the pattern from a drawing, are the parts that software handles most cleanly. CAD and CAM tools have done this for years. By share of task time, the AI-does group sits at 5%.
Assisted work is wider. Reading specifications, planning cut sequences, and sizing stock all go faster with a digital model in front of you, but a person signs off on each call. The assisted share comes to 19%. Overall coverage, our measure of how much task time AI can handle today, reads 15 out of 100; the short explainer on how coverage is measured sets out the method.
Then the bench itself. Shaping and assembling pattern sections, repairing and reworking existing patterns, and checking fit and finish before the pattern goes to the foundry stay with people. The human group holds 76% of task time. These are not tasks waiting on better models. They are tasks waiting on hands.
What has actually been tested
Not much, and that is the honest answer. The evidence grade for this job is D, our lowest tier, which means no study has put an AI system against a qualified wood patternmaker on this job’s real tasks. Because of that, we publish no quality-parity number here. A grade with no direct test is a gap, not a verdict.
What would settle it: a timed trial on pattern production from a cast-part drawing, measured on dimensional accuracy, draft, and whether the finished pattern draws cleanly in molding sand. Until something like that exists, the parity question stays open. Our page on how parity is graded explains why we refuse to put a number on untested work, and the broader scoring method covers the rest.
When the picture could shift
Most likely after 2046 (8 in 10 of our scenarios). For what that window measures and how we build it, see the note on replacement-year estimates.
Two things could pull the date closer. Cheaper five-axis CNC routers and large-format 3D printing move more pattern geometry straight from file to finished form, skipping the bench for simple shapes. And foundries that already model parts digitally may order printed sand molds instead of wooden patterns at all, which changes the demand rather than the task.
Two things hold it back. The physical share of this job is high, and the robotics available to it is fixed automation: machines that do one programmed path well and need re-rigging for every new job. That economics fits long runs, not the one-off and repair work that fills a pattern shop. And the trade is shrinking slowly on its own, with BLS projecting a 2% decline in employment between 2025 and 2035 (BLS, 2025). Few new entrants means few shops investing in new equipment either way. If you want to see how physical work resists automation generally, the guide on robots and physical jobs covers the cost side.
What to do: learn the CAM side of your own shop’s machines, so the file work stays yours rather than moving to an engineer two states away.
How to stay needed in a pattern shop
Lean into the three tasks that nothing has touched. Repair and rework of existing patterns, where the original drawing is often missing. Fit checking and correction against the casting that comes back. And the draft, fillet, and finish decisions that decide whether a pattern survives a hundred pulls or ten.
Two skills raise your floor. CAD and CAM fluency, so you can take a customer’s model and drive the router yourself. And foundry literacy: knowing how sand behaves, why a casting pulled short, and what to change in the pattern rather than the process. Both make you the person who closes the loop.
Nearby work is worth comparing if you want options. Model Makers, Wood is the closest match in both tools and judgment. Patternmakers, Metal and Plastic covers the same job in different stock. Cabinetmakers and Bench Carpenters is the larger trade with the most transferable bench skills. You can also browse the wider woodworking occupations, see how the rest of manufacturing scores, or put two trades side by side on the job comparison tool. The list of jobs that mostly need a person shows where hands-on work sits against the rest.