Will AI replace Drafters, All Other? 28/100

COOKED

This residual drafting category — the specialists who don't fall under architectural, civil, mechanical, or electrical drafting — spends most hours converting engineers' sketches, specs, and marked-up redlines into dimensioned CAD models, sheet sets, and BOMs. Parametric CAD, generative layout tools, and AI that reads specs and produces drawing packages already handle the bulk of that translation work, and the drafter carries no stamp: a licensed PE or architect signs the drawing, so no regulatory shield protects the role. What survives is the coordination tier — clash detection, as-built field verification, tribal knowledge of a shop's fabrication limits, and cleaning up the AI's plausible-but-unbuildable geometry.

10-year outlook: Headcount in this small residual category keeps shrinking as spec-to-drawing generation matures; the remaining jobs look less like drafting and more like model coordination and field verification.

28/100 risk resistance · ±3 on re-score

US employment, 2019–2025-0.9%

14,60014,470 workers

Roughly flat across the period, with year-to-year wobble.

Median pay $52,830 → $63,520 -3.8% in real terms (nominal +20.2%, less ~25% US inflation over the period)

The job count is not the verdict

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

-6.9%

Percentage only. The projection counts a different population from the 14,470 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.

Exposed, and shrinking

Both signals point the same way: the tasks are largely automatable and the BLS projects -6.9% by 2034. This is the case where the score and the forecast agree, and it is the one worth taking seriously.

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.

~1,300 openings a year on average, including replacing people who leave.

One email if this score changes. Watch as many occupations as you like from the same address — no account, and nothing is sent on a schedule, only when a verdict actually moves.

Also known as — 24 job titles this covers

Titles reported by people doing this work, from the US Department of Labor's O*NET survey. If your job title is here, this page is about your work even though the name doesn't match.

TracerDetailerDraughtsmanJob CaptainHull DrafterChief DrafterForm DesignerDetail DrafterDrawing TracerMarine DrafterSteel DetailerDrawing CheckerBlueprint TracerComputer DrafterDrafting DetailerDrafting EngineerLaboratory WorkerAutoCAD TechnicianBlueprint EngineerCommercial DrafterDrafter ApprenticeGeological DrafterDrafting Layout ManEngineering Drafter

This is a catch-all code, not a single job

The BLS uses Drafters, All Other for work that doesn't fit any named occupation, so it covers roles that have little in common with each other. Two consequences worth knowing before you read anything below:

The score is an average over dissimilar work. Your own role could sit well above or below it, and the five dimensions describe the bucket rather than any one job inside it.
We've withheld the transfer routes here. They are computed from skill-profile overlap, and a bucket's profile is an average of unrelated jobs — which produces confident-looking matches that are meaningless. Better to show none than to point someone at the wrong career.

If a more specific occupation on the register describes what you actually do, that page is the one to trust.

Score — 28/100 resistance

Holding it up: task resistance (7/20). Weakest point: liability shield (3/20).

Five dimensions, 0–20 each, summed. Higher means more protected. The arithmetic is shown so you can check it: 7 + 6 + 3 + 5 + 7 = 28. · Scored 2026-08-11, and re-examined when evidence accumulates rather than on a schedule.

Task resistance 7/20

Mixed — a routine tier and a judgment tier At 7 the score credits the part of your week that isn't translation — reconciling a fabricator's tooling limits against a designer's geometry, chasing down which revision of a redline is current, catching that the AI-generated layout put a weld seam where nobody can reach it — but the dimensioning, sheet-set generation, view creation, BOM extraction, and revision-cloud markup that fills most timesheets is already handled by parametric templates and spec-reading tools, which is why it sits at 7 rather than 12.

Embodiment 6/20

Some physical or field component A 6 rather than a 2 reflects the field trips: as-built verification with a tape and laser measure, walking a shop floor or plant to confirm a pipe run or panel location that the model got wrong, and occasionally handling physical samples or legacy mylars — real but episodic, since the model tree and the monitor are where the hours actually go.

Liability shield 3/20

No licence, no signature requirement 3 is not zero only because some employers want ASME Y14.5 GD&T certification or a Certified SolidWorks Professional credential on the resume; nothing in state engineering practice acts requires a drafter to hold anything, and the seal that makes a drawing legally issuable belongs to the PE or RA who reviews it, so no statute makes you the last line.

Trust premium 5/20

Anonymous artifact production 5 acknowledges that the engineer who has worked with you for six years hands you a napkin sketch instead of a written spec because you already know the house standards — but the deliverable is a DWG or STEP file that leaves the building with someone else's name in the title block, and it gets judged on whether it builds, not on who drew it.

Judgment & accountability 7/20

Meaningful discretion 7 recognizes real calls — picking tolerance stack-ups, deciding how to break a complex assembly into sheets, flagging an interference the designer missed — but those decisions are bounded by ASME Y14 standards, company drafting manuals, and a checking/approval loop where the engineer of record absorbs the consequence of a bad one.

Confidence: high · reviewed 2026-08-11 · how scoring works

How to future-proof this job

Move into BIM/CAD coordination: own clash detection, model federation, and the drawing-control system rather than sheet production.
Get on-site: build a reputation for as-built laser scanning and field verification that turns messy reality into trustworthy models.
Learn the fabrication side — welding, machining, or assembly constraints — so you are the person who catches unbuildable AI-generated geometry.
Stack toward the licensed tier: pursue engineering technology credentials or an EIT path so you eventually sign work instead of preparing it.

All 35 skills ranked by how many jobs they open →

What would move this back up — beyond any one person

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 41/100 — EXPOSED.

4 specific changes that would raise this score

already happening task resistance +4

Task-mix shift to the coordination tier: if the routine redline-to-sheet translation is fully automated, what remains is BIM/model clash resolution, verifying AI geometry against a specific shop's fabrication limits (press-brake radii, weldment access, stock sizes), and reconciling as-built field conditions with the model. This tier is genuinely resistant because the constraints live in undocumented shop practice, not in any spec the model can read. Watch for job postings retitling to 'BIM coordinator' or 'design technician' with fabrication-review duties.
plausible judgment accountability +4

If the drafter becomes the designated model checker of record in a formal BIM Execution Plan (AIA G202 / ISO 19650 information-manager role), owning documented clash-resolution decisions and named as Model Coordinator on the contract, the role owns consequential calls under ambiguity rather than transcribing them.
plausible embodiment +3

If as-built verification and field measurement become a routine assigned duty — laser scanning / reality capture on site, then registering point clouds against the model — the role acquires an unpredictable-environment component. Already visible where contractors run Matterport/Leica scan-to-BIM workflows in-house rather than subcontracting.
unlikely liability shield +2

Narrow route only: state PE boards or an employer's professional-liability insurer requiring a named, qualified human to certify that AI-generated drawing packages were independently reviewed before the PE seal is applied — analogous to the responsible-charge review language already in NCEES model rules on delegated work. This creates a documented human checkpoint, but the stamp and personal liability still sit with the PE, so the gain is small.

The limit. No plausible route to a higher trust premium: drafting is a purchased intermediate output inside an engineering deliverable, and no buyer pays a premium for human-drawn sheets. Even with every lever, the ceiling is roughly mid-40s — the role's structural problem is that the stamp belongs to someone else, and no realistic rule change transfers it to the drafter.

These are conditions, not forecasts — what would have to happen, not what will. Specific rules, cases and bills are named so you can go and check whether they exist and where they stand; verify before relying on any of them. Nothing here is legal or financial advice.

Where this work is, and what it pays there

BLS metro figures for 82 areas. The verdict above does not change by city — the rubric judges what the work involves, not where it happens — but pay and headcount do, and the national median hides a very wide range.

Most of these jobs

Dallas-Fort Worth-Arlington, TX	700	$61,200 -4%
Houston-Pasadena-The Woodlands, TX	690	$60,820 -4%
Los Angeles-Long Beach-Anaheim, CA	690	$73,840 +16%
New York-Newark-Jersey City, NY-NJ	550	$81,220 +28%
Portland-Vancouver-Hillsboro, OR-WA	450	$62,610 -1%
Atlanta-Sandy Springs-Roswell, GA	380	$62,500 -2%
Phoenix-Mesa-Chandler, AZ	330	$70,320 +11%
Seattle-Tacoma-Bellevue, WA	330	$72,990 +15%

Best paid

San Jose-Sunnyvale-Santa Clara, CA	80	$90,850 +43%
San Francisco-Oakland-Fremont, CA	170	$82,360 +30%
New York-Newark-Jersey City, NY-NJ	550	$81,220 +28%

Percentages are against this occupation's national median of $63,520. Counts are jobs in that metro, not vacancies. Metros where the BLS suppressed the cell are absent rather than shown as zero.

Who is actually doing this — nobody, on the record

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 28. 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.

Read that as a gap in the reporting we can see, not proof of absence — the dispatch runs on English-language feeds and misses plenty. If you know of a case, tell us, or add a field report from inside the job.

Quick take — do you do this job?

Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.

Self-reported and unverified — a sentiment signal, not a survey. One response per person per occupation; you can change your answer.

Field reports — what people say has changed

No field reports yet. A written account takes a paragraph rather than a tap, goes to an editor before it appears, and is the one thing on this page the rubric cannot produce on its own.

Watch this verdict

Kept current

The dispatch runs daily. 711 reported cases on file, 357 of them things a named organisation actually did. Newest landed yesterday. See it · RSS
Scores move on evidence, not on a clock. Re-scoring the same occupation drifts a couple of points on its own, so a scheduled refresh would mostly republish our own noise. An occupation is re-examined when deployments accumulate against it — which is why a verdict changing here means something. Last register revision 2026-08-11.
Employment updates once a year, when the BLS publishes. Every occupation carries its own 2019–2025 series.

Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.

Cooked Index — Drafters, All Other: COOKED, 28/100.
Scored 2026-08-11 · printed from https://cookedindex.com/jobs/drafters-all-other/
Verdicts are re-examined as evidence accumulates, so this sheet goes out of date. Check the page above for the current score, the reasoning behind each dimension, and what would move it.

Drafters, All Other