Architectural and Engineering Manager

Architectural and Engineering Manager scores 27/100 on this site's structural exposure measure (Medium confidence), which falls in the Higher Risk band. Under this site's research checklist, a low score is offset only by a legal wall or a physical requirement, and neither is documented for this role. BLS says these managers typically do not need a license. The licenses nearby belong to the people they manage: every state licenses architects, and some engineers hold a Professional Engineer (PE) license.

A manager may hold those credentials, but the job does not require one, so no credential is tied to the managing itself.

The work is also not physical. BLS says most of these managers work in offices, and O*NET's 19-task list is dominated by verbs such as prepare, review, present, assess, approve and negotiate: budgets, bids, cost estimates, contracts, design-change approvals, client presentations and staffing. That is the kind of language-and-documents work both exposure measures rate highly.

The usual defense, that managing people is inherently human, is plausible, but this site's checklist does not accept it as sufficient without a legal or physical wall behind it.

What AI is actually doing in this field is narrower than the score implies. Procore, the construction-management platform, lists agents for RFIs, submittal review, contract review and daily logs. Its AI page says agents never take action without sign-off and that the team makes the final call. Those tools take over drafting and checking, the paperwork a project engineer or manager used to do or supervise. They do not replace the person who decides.

Adoption is partial. A Bluebeam survey of about 1,000 architecture, engineering and construction professionals (reported by ASCE in December 2025) found 27% using AI in their operations. The AIA's August 2026 Architecture Billings Index special question found 48% of responding firm leaders say their firm is engaging with AI, and only 7% have a formal AI training program. The two surveys define use differently, so they should not be compared as one trend.

We searched for disconfirming evidence, meaning a named engineering or architecture firm cutting engineering or project managers because of AI, and found none. The nearest hard data is from software. The LeadDev Engineering Leadership Report 2026, based on 600 engineering leaders surveyed between 31 March and 16 April 2026, found 22% reporting a decline in managerial jobs and 19% an increase.

That is a software-team survey, not an architecture or engineering one, so it is a warning sign rather than evidence about this occupation.

One rule does protect the licensed engineers these managers lead. In NSPE Board of Ethical Review Case 24-2 (July 2024), an engineer who sealed AI-drafted designs after only a cursory review was found not to have maintained responsible charge, and the Board said AI output needs at least the same scrutiny as human work. That binds the engineer who seals, not managers as a class, and the case is now more than two years old.

BLS projects 6% employment growth for the occupation from 2025 to 2035, faster than average: about 12,200 jobs on a 2025 base of 220,900, with roughly 14,900 openings a year. That is a headcount projection. It does not describe how the tasks inside the job are changing.

The score and the field data therefore point in different directions: high exposure on paper, no documented displacement, and growing projected headcount. The checklist says to apply its rubric rather than override it, and to flag the score for review, so the verdict is At risk with Medium confidence. Read it as a statement about task exposure for the occupation, not a forecast of job losses.

Our own inference, not a sourced finding, is that the likelier near-term change is a shift in how managers spend their time and pressure on how many managers each team needs, rather than elimination.