The systems engineering talent shortage is not a new problem. The shortage has been identified in workforce studies from INCOSE, AIAA, and the Defense Acquisition University for over a decade. But the gap between demand and available experienced practitioners has widened, and the structural factors driving it aren’t resolving.
At the same time, the first generation of AI-assisted systems engineering tools is changing what’s possible with the engineers teams already have. The combination of these trends is reshaping how engineering organizations are thinking about systems engineering workforce strategy.
The Shape of the Gap
The shortage has two components that are often conflated but have different causes and different solutions:
Retirement-driven loss of experience. The generation of systems engineers who built their expertise on Cold War and post-Cold War programs — aircraft, missiles, satellites, submarines — is retiring. Their institutional knowledge about how complex systems fail, what requirement patterns lead to integration problems, and how to navigate certification processes is leaving with them. You can hire a new graduate with a systems engineering degree, but you can’t replicate 30 years of program experience quickly.
Growth-driven demand increase. Defense program activity has increased significantly. The commercial aerospace backlog remains large. The electrification of automotive and industrial products has created new demand for systems engineers in sectors that didn’t historically employ them heavily. Autonomous systems, industrial robotics, and medical devices are all pulling from the same limited pool of experienced systems engineers.
The combination means that organizations are running programs with systems engineering teams that are more junior than historical norms, more stretched across concurrent programs, and less able to rely on institutional knowledge that used to be available informally.
What AI Tooling Is Changing
The most immediate impact of AI-assisted requirements tools on the talent gap is leverage: letting experienced systems engineers cover more surface area, and letting less experienced engineers produce work of adequate quality faster.
Automated quality checking means a senior systems engineer reviewing requirements can cover more requirements per hour because the obvious quality issues — ambiguous verbs, missing measurability, compound requirements — have been flagged before they arrive. The senior engineer’s time is spent on substantive issues, not formatting and phrasing.
AI-assisted decomposition means a systems engineer working on requirements decomposition has a starting point rather than a blank page. First-pass child requirements generated by an AI model, reviewed and refined by the engineer, take less total time than the engineer generating everything from scratch. The quality of the AI first-pass matters — garbage starting points waste more time than they save — but mature tools are producing starting points that experienced engineers find useful.
Impact analysis acceleration means that when requirements change — which they do constantly in active programs — understanding the downstream impact no longer requires a senior engineer spending a day manually traversing the traceability model. The impact set is a query. The senior engineer reviews and validates the impact set rather than assembling it.
Knowledge capture through structure means that when an experienced systems engineer makes a requirements decision — allocates a performance budget a certain way, decomposes a requirement in a specific structure — that decision is captured as structured data in the model rather than as prose in a document that gets filed and forgotten. Decisions in structured models are more accessible to junior engineers trying to understand why the requirements are the way they are.
Rethinking Where Expertise Is Required
The talent gap is pushing organizations to explicitly ask which systems engineering activities require deep expertise and which require trained capability with AI assistance.
Activities where deep expertise remains irreplaceable:
- Requirement capture from operational stakeholders who don’t speak systems engineering
- Performance budget allocation where system behavior emerges from component interactions
- Failure mode analysis for complex systems
- Navigating certification requirements with regulatory authorities
- Making architectural decisions where multiple approaches are defensible and the choice has long-term consequences
Activities where AI-assisted tooling with less senior engineers is increasingly viable:
- First-pass requirements authoring from a brief or stakeholder statement
- Decomposition of well-understood system types
- Traceability maintenance and coverage monitoring
- Requirements quality review and documentation
The teams that are navigating the talent gap most successfully are doing explicit workforce planning around this distinction — deploying their most experienced systems engineers on the activities where experience is irreplaceable and using AI-assisted tooling to extend their capacity on the activities where it can substitute.
Knowledge Transfer as Strategic Priority
One response to the retirement wave that’s gaining attention: treating knowledge capture from experienced engineers as an explicit program activity, not as an outcome of good documentation practice.
The specific form this is taking in organizations with mature AI tooling:
Structured interview capture. Working with experienced engineers to capture their knowledge in forms that translate into structured requirements model patterns — not prose documents but actual requirements hierarchies, allocation patterns, and traceability structures that encode how they would approach decomposing a system of this type.
Decision documentation in the model. Requiring that allocation decisions, architectural choices, and requirements derivation rationale be documented in the model with explicit rationale rather than implied by the structure. This makes experienced engineers’ reasoning accessible to engineers who weren’t present when decisions were made.
Mentoring through tool use. Pairing junior engineers with senior engineers in the requirements model environment, where the senior engineer can show the junior engineer how to use the AI assistance effectively — when to accept AI-generated output, when to reject it, how to review decompositions for completeness.
Process pattern libraries. Capturing common requirements patterns for recurring system types — power distribution subsystems, RF communication systems, software safety monitor architectures — as reusable starting points that encode the knowledge of engineers who have built these systems before.
This kind of deliberate knowledge transfer through structured tooling is a different approach from the traditional apprenticeship model, which required junior engineers to be physically co-located with senior engineers for years. It’s not a complete substitute — tacit knowledge still transfers through relationship and experience — but it’s substantially better than the alternative of letting retirement take the knowledge with it.
The Honest Assessment
AI tooling is not going to close the systems engineering talent gap. The gap is structural, the retirement wave is real, and the demand growth is not slowing. Hiring, training, and retaining experienced systems engineers remains the foundational workforce strategy.
What AI tooling does is change the leverage ratio and the substitution frontier. With good AI-assisted tools, an organization can do more systems engineering work with the same number of experienced engineers, and can execute more of the work that requires less expertise with less experienced engineers.
Organizations that are combining deliberate AI tooling investment with deliberate knowledge transfer programs are building more resilient systems engineering capacity than those treating tool selection and workforce development as separate problems. The teams that will perform best in an era of constrained systems engineering talent are the ones that are taking both seriously.