SECC

Origin

Began in MZI's internal ideation practice; developed into a GMU capstone MVP.

External Links

• Peer-reviewed paper (ITEA Journal)
• Product site

The Full Story

SECC started as a question inside MZI's own engineering ideation practice: could an AI system genuinely help with the document-heavy assurance work that eats up so much of a systems engineer's time on federal integration efforts, without pretending to replace the judgment that work actually requires.

That question turned into a sponsored capstone project with George Mason University's Systems Engineering and Operations Research program. A graduate student team built the MVP: an AI-powered assurance agent that reviews systems engineering artifacts for the kinds of inconsistencies and gaps that are easy to miss and expensive to catch late.

The GMU partnership is where SECC started, not where it stayed. Since the MVP came together, we've taken it beyond the classroom and demoed it as a working prototype to prospective customers. The academic work also produced something useful in its own right: a peer-reviewed evaluation, published in ITEA Journal (Volume 47, Issue 2, June 2026 DOI: 10.61278/itea.47.2.1007), that tested the MVP against three curated document sets with deliberately seeded issues. It met an 85 percent detection threshold on those issues, cut review time by more than half, and pointed toward substantial cost reduction relative to fully manual review, based on a parametric cost model.

That evaluation was a controlled test on seeded material, not a production validation, and we're upfront about that. A formal precision and recall study still needs to happen before we'd make claims about real-world performance. What's true today: SECC is a working prototype with real evidence behind it, already being shown to real customers, with more validation work ahead before it's something we'd put in front of an operational program.