Why most of the day still sits with a person
The cutting cycle on a drill press or boring mill was automated long before anyone asked will AI replace drilling and boring machine tool setters, operators, and tenders. A program repeats the hole pattern all shift. What software has not taken over is the work around the cycle: clamping an awkward casting square in the fixture, choosing the bit and feed for cast iron rather than acrylic, hearing chatter start, pulling a dull tool before the hole drifts out of tolerance, and checking the bore with a gauge and trimming the offset.
That is hands, eyes and local knowledge. The parts change, the fixtures are shop-specific, and the tolerance call often comes down to feel plus a measurement. Our robotics read puts most of this job in physical work, and the automation route is fixed automation: dedicated equipment bought for one part family, not a general-purpose robot that learns a new shop in an afternoon.
Fixed automation pays off on long runs of the same part. Many shops running drilling and boring work are smaller job shops with frequent changeovers, short batches and mixed materials. Every changeover is setup labor, and setup is the part a machine cannot buy its way out of. You can see how the share of automatable task time is built on our coverage method page.
What the software runs, what it assists, what stays with you
Only a slice of the task time can run without a person today: 0%. That slice is the repeatable end of the job, where the program drives a known hole cycle and the control logs run data, spindle load and cycle counts without anyone standing there.
A similar slice is assistance rather than replacement: 7%. Here the tool suggests speeds and feeds, flags a spindle-load pattern that looks like a dull bit, or pulls dimensions out of a drawing so you do not retype them. The operator still makes the call and still turns the wrench.
The rest is work that needs a person at the machine: 93%. Setting and squaring fixtures, loading and unloading parts, swapping and touching off tooling, first-article inspection, deburring, and deciding whether a part is scrap or can be saved all sit in that group. The task list above shows which jobs fall where.
What has been tested, and what has not
Our evidence grade for quality against a qualified operator is D. In plain terms, there is no direct published test of an AI system against a trained setter on this job, so we give no parity number. We will not guess one.
What would settle it is a timed trial on real machines: mixed parts and materials, setup from a print, first-article inspection, tolerance held across a batch, with scrap rates and rework reported. Until something like that is published, the honest answer is that the outcome is unmeasured, not favorable or unfavorable. Our rules for grading that kind of test are on the quality parity page, and the wider scoring approach is set out in our methodology.
The labor market numbers are clearer. The Bureau of Labor Statistics counts about 4,680 people in this occupation, with median pay of $49,080 and a projected change of -9.5% from 2025 to 2035 (BLS, 2025). That decline is driven by machine consolidation and offshoring pressure more than by anything new in software.
When the timeline could move
Most likely after 2046 (8 in 10 of our scenarios). The way that window is built, including the median and the spread, is explained on the replacement year page.
Two things could pull it earlier. Cheaper part-handling cells, where a robot arm loads and unloads a drilling machine reliably across more than one part shape, would eat into the loading tasks. And a sharp drop in the cost of vision-based in-process inspection would move measurement work onto the machine. The cost panel above shows how far apart equipment and labor costs sit today.
Two things hold it back. First, fixturing: every new part needs a physical setup, and that is design and judgment work per job, not a software update. Second, batch size. Dedicated automation only earns its keep on volume, and a lot of this work is short runs. Scrap liability matters too, since a hole drilled in the wrong place on an expensive casting is money gone.
Good to know: fewer entry-level machine-tending openings is the more likely pattern here, because one person increasingly watches several spindles.
Staying needed on the shop floor
Lean into the parts of the job that travel with you. Setup and fixture design is first: being the person who can hold an odd part rigid and repeatable is hard to buy. Metrology is second: bore gauges, micrometers, surface finish checks and first-article inspection against a print. Troubleshooting is third: reading a bad hole and knowing whether it is tool wear, runout, coolant or a soft fixture.
Two skills are worth real time. One is CNC programming and editing, including G-code, offsets, tool tables and probing routines, so you set the machine rather than only tend it. The other is reading geometric dimensioning and tolerancing well enough to argue a print with an engineer. Both move you up the value chain inside the same shop.
If you want to look sideways, close neighbors include lathe and turning machine operators, milling and planing machine setters and computer numerically controlled tool operators, which is the usual step up for setters who learn programming. The metal and plastic workers family page shows the rest of the group, and the manufacturing sector page puts it in context with the industry around it.
From there, two things are worth checking: put this job next to a nearby one on our compare tool, and see where machining roles land among jobs expected to shrink before you commit to a training plan.