The Materiel Command’s Engineering Division (A4/10), a key logistics and engineering directorate of the U.S. Air Force, has adopted Box as an AI‑ready, FedRAMP High and IL4‑authorized Intelligent Content Management platform. This strategic technology move enables secure collaboration and will accelerate key workflows for federal government needs.
This is important for engineers in A4/10, because modernizing aircraft and ensuring mission readiness depends on mastering terabytes of digital documents. Ms. Donna Stacy, a Principal Systems Engineer, leads a team responsible for everything from developing engineering policy to reviewing new technology proposals. These are tasks that require synthesizing dense Department of Defense instructions, multi‑volume manuals, and lengthy reports.
Legacy content management created constant friction with these tasks. Converting to Box as an AI-ready platform enables Stacy’s team to:
- Reduce technical proposal review time from ~2 hours to a fraction of that, saving approximately 30–40 hours every two months across ~20 proposals.
- Create custom Box AI Agents (using Box AI Studio) that provide strong, editable starting points for complex documents like JICDs, charters, and white papers
- Ensure secure, verifiable outputs rich with citations and true to FedRAMP High and IL4 authorization
Tedious reviews and information overload slowed down work
When Stacy’s team was tasked with creating a Joint Initial Capabilities Document (JICD) for a new maintenance program, the limitations of legacy systems became clear. The document is large, tightly coupled to guidance from the Joint Capabilities Integration and Development System (JCIDS), and requires alignment with strategic plans and joint warfighting concepts. Manually finding, comparing, and aligning source material was slow, error‑prone, and frustrating.
Specific challenges included:
- Tedious document creation across disparate, unstructured sources that slowed synthesis and increased errors
- Time‑consuming reviews where a single technical proposal could take hours and interruptions multiplied the cost
- Conflicting guidance, acronyms, and fragmented records complicated federal records management and FOIA‑ready outputs
The team needed a consistent way to automate heavy document analysis and creation so engineers could focus on high‑value, mission‑critical work.
Box AI Studio and custom agents
Using Box AI Studio, Stacy built custom Box AI Agents tailored to the most common, time‑consuming tasks. Agents were trained on relevant content and configured to respect Zero Trust and NIST 800‑53 oriented controls. Key agents included:
- Proposal reviewer: Analyzes technical proposals against criteria, tabulates financial plans, and summarizes key points
- Policy analyst: Reviews policy documents, distinguishes mandatory requirements from guidance, and identifies bottlenecks
- Contribution statement writer: Drafts performance appraisals by synthesizing activity reports and position descriptions
All agent outputs include a built-in human‑in‑the‑loop review step and provide bulleted citations (with page/section) to make results verifiable and FOIA‑traceable. Role‑based access controls are respected during queries, preserving secure, compliance‑aware results.
The proposal reviewer agent already saves teams 30 to 40 hours a month on roughly 20 proposals.
Faster reviews, more innovation, and secure verifiable outputs
The agency has accrued significant time savings and improved outcomes from using these custom Box AI Agents. The proposal reviewer agent already saves teams 30 to 40 hours a month on roughly 20 proposals. Agents also provide a strong starting point for documents like charters, white papers, and JCIDS that engineers can then use their subject-matter expertise to quickly edit.
“As an engineer, I don’t like writing, but Box AI helps me coalesce my thoughts into something very well written,” Stacy says. “One of my biggest problems is that I get interrupted. Having to stop and start reviewing documents is the biggest time sink. Pulling the relevant information out quickly helps me get things done faster.”
The fact that the agents have been trained to include citations with their outputs, combined with the Box platform’s FedRAMP High / IL4 posture and RBAC awareness, give the agents credibility. This, in turn, encourages adoption throughout the agency.
Everyone was shocked at how good it was. And Box helped me pull it all together.
Relying on Box AI to adapt to constantly changing guidance and requirements
Box has made work easier and outputs more reliable in other ways, as well. When JCIDS guidance changed, rather than restarting work from scratch, the team was able to use Box to evaluate updated instructions, pinpoint changes, and validate requirements against source material. That made it much easier to reformat and revise the JICD quickly and keep it aligned to evolving requirements.
“I had an assist from Box AI for the first draft,” Stacy recalls. “It was the most thorough JICD I’ve ever seen. Everyone was shocked at how good it was. And Box helped me pull it all together.”
In another instance, the main CBM+ (Condition Based Maintenance Plus) capabilities document required consolidating requirements from many prior policies and procedures. By bringing all relevant information into Box, Stacy created a roadmap and suggested maintenance procedures that unblocked the stalled project.
Box AI has been one of the most successful technologies that I’ve seen in a long time
The next steps with custom Box AI agents
Embedding AI into secure, FedRAMP High / IL4‑authorized content workflows moved Stacy’s team from reactive firefighting to proactive decision‑making. Teams requested licenses, built their own custom agents, and replicated agent templates across offices to capture consistent time savings. A4/10 is now expanding pilot users, aligning agent behavior with rigorous security and compliance controls, and rolling out proven templates to amplify organizational capacity.
“Box AI has been one of the most successful technologies that I’ve seen in a long time,” Stacy says. While she advises reviewers to always verify outputs (“maybe it’s 80% of the way there”), that 80% — backed by verifiable sources and secure access controls — was enough to drive adoption and free subject‑matter experts to focus on higher‑value technical and strategic work.
By centralizing documents in Box Hubs and deploying custom, compliance‑aware AI agents, A4/10 accelerated document review, improved traceability, and reclaimed hours for strategic work — enabling faster, more confident decisions that help keep the Air Force ready for combat.
Read more government customer stories on Box.
PA Release: AFMC-2026-0023
