EXPOSED
Much of the work — irrigation and drainage calculations, CAD drafting of equipment and storage structures, sensor/yield data analysis, grant and permit paperwork, feasibility write-ups — is exactly the structured technical output AI now drafts competently. What holds is the field half: walking fields and feedlots to diagnose why a drainage plan or manure-handling system actually failed, commissioning machinery, and taking responsibility for designs that touch food safety, water rights, and livestock welfare. PE licensure applies to stamped structural and water-resource work but a large share of ag engineers never stamp anything, so the shield is partial.
Mixed — a routine tier and a judgment tier. Sizing a tile drain lateral, running NRCS curve-number runoff models, drafting a grain bin in SolidWorks, and cleaning telemetry from soil-moisture probes are all rule-bound enough for software to carry, but nobody automates the trip out to the failing lagoon berm or the retrofit of a combine header on a specific farmer's ground — which splits the day roughly in half and puts this at 11 rather than up with field-only trades.
Some physical or field component. You are on site for pump commissioning, laser grade checks, feedlot and manure pit inspections, and machinery trials in mud and dust — but that is episodic fieldwork bracketing weeks at a workstation doing hydraulic calcs and drawings, so 11 rather than the 15+ of an equipment operator who never leaves the field.
Certification preferred, not legally required. The 10 reflects a real split: water-resource designs, retention structures, and grain and livestock buildings that get stamped carry PE personal liability under state engineering practice acts, while ag engineers in machinery R&D, extension, or ag-tech firms often work under the industrial exemption and never seal a drawing in their careers.
Some relationship component. Repeat work for the same producers, co-ops, and drainage districts means growers ask for you by name after your last plan held through a wet spring, but the deliverable is still a design and a report someone else could have produced — the relationship shortens the sales cycle, it isn't the product.
Meaningful discretion. You decide design storm return periods, setback distances from wells, ventilation rates that determine whether livestock die in a heat event, and whether a lagoon liner spec is adequate under state nutrient-management rules — genuine consequential calls, but bounded by NRCS standards, ASABE specs, and permit conditions rather than made in a vacuum, which is what keeps it at 12.
Has AI actually changed your work?