By Michael MacFadden, CTO of Sigma Defense
Imagine a scenario in which a U.S. military unit operating in a heavily jammed, contested environment has its primary routing software crash just as it must resupply a forward air defense battery before an anticipated missile attack.
An agentic software development system running on secure local infrastructure rapidly analyzes the failure, generates and tests a patched routing application using degraded but available data sources and deploys it in time for commanders to reroute convoys around disrupted roads and known threat areas.
While this happens mostly autonomously, there is still human oversight through the chain of events. That speed preserves the battery’s readiness, prevents a gap in air defense coverage, and improves the mission’s chance of success.
Agentic software development is the next step in democratizing battlefield capability, as it lowers the barrier between operational needs and deployable software.
Future conflicts will favor those who convert frontline needs into software solutions, decision tools, and workflows faster than adversaries.
Agentic development makes that possible by enabling smaller teams, nontraditional developers and edge units to participate in building and refining mission software under governed conditions.
Agentic coding closes the mission development gap
The traditional defense software model is too centralized and too slow for dynamic, tactical environments. Units in the field often identify urgent needs long before contracting, development, testing and deployment cycles can respond.
Historically, these problems rarely justified the development of new software. The acquisition process was too slow, and the development cost was too high for localized issues affecting only a small number of operators. Even when a solution was technically feasible, it often arrived too late to matter.
Agentic development bridges operational needs and software solutions by combining human input in operational language with intelligent systems for rapid prototyping, workflow automation, code testing, and documentation.
This combination produces quicker adaptation at the tactical edge and less dependence on scarce elite technical talent, shortening the time from identifying the problem to creating a usable tool.
Looking to the past for institutional inspiration
This kind of grassroots adaptation is not unfamiliar in military history. During the early phases of the Iraq War, improvised explosive devices exposed a dangerous vulnerability in unarmored vehicles. The industrial base eventually responded with factory-armored platforms, but that response took months and years.
While they waited, soldiers and Marines improvised by welding additional steel plates onto vehicles and attaching body armor to doors. While these solutions may have been inelegant, they were operationally valuable.
Imagine warfighters having the ability to create mission software in much the same way they once improvised vehicle armor. A crew aboard a naval vessel might build a workflow automation tool that reduces administrative burden.
A forward unit could rapidly integrate a newly deployed sensor into a situational awareness dashboard. Analysts in a tactical operations center might construct visualizations that reveal patterns in operational data.
These tools may not require formal acquisition, but each can deliver operational gains. When software creation becomes nearly costless, access to building tools is a critical factor. The widespread availability of agentic development tools empowers frontline units to respond more quickly to evolving mission needs.
Challenges that lie ahead
Enabling this model introduces real challenges that must be addressed thoughtfully.
Many modern agentic coding tools rely on cloud infrastructure, while tactical environments often operate in disconnected or classified conditions. Development tools must therefore be able to operate locally.
Training matters. Agentic systems lower the barrier to software creation but do not eliminate the need for skill. Warfighters must understand the problems they are solving and how to guide these tools effectively.
Cybersecurity and accreditation also remain serious concerns. Introducing software created in tactical environments requires new approaches to governance and security.
Beyond that, robust automated testing and guard rails are required to make sure software quality and security are maintained. The changes made rapidly in the field can still be shipped back to high performing software teams to harden those changes, while still allowing rapid response in the field.
Rethinking the software factory
Some of the challenges outlined above can be solved by changing the way the DoD thinks about software factories.
Today, most defense software factories are centralized organizations designed to support enterprise capabilities and major programs. They provide development infrastructure, testing pipelines, security tooling, and deployment environments.
In the era of personal software, we will likely need a complementary model operating closer to the operational edge.
Tactical software factories could provide warfighters with tools, training, and pre-approved infrastructure while embedding automated security and testing controls directly into the environment.
Agentic development shifts software creation to users and small teams, enabling rapid, tailored solutions to operational needs without reliance on slow, centralized processes.
In the commercial world, this shift is ushering in the era of personal software. In the military domain, it opens the door to grassroots innovation at the tactical edge. Software mobility determines where software can run.
In the era of agentic development, we must ensure mobility in where software can be built as well. The next operational advantage may not originate from a centralized program office or a large development team. It may come from the operator closest to the problem.
##
ABOUT THE AUTHOR
Michael MacFadden is a proven technology leader with over 20 years of experience of delivering innovative solutions to DoW customers.
Michael brings multi-disciplinary engineering expertise, strong technical leadership, and a passion for injecting new technologies and innovation into the defense market. As CTO, Michael is responsible for seeking out new technologies to solve complex customer problems across multiple technology domains. He combines strong technical ingenuity and business acumen skills to help Sigma Defense deliver new solutions and services that address customer’s unique requirements.
Michael joined Sigma Defense in January 2022 through the acquisition of SOLUTE Inc, where he also served as Chief Technology Officer. During his 11-year tenure at SOLUTE, Michael was responsible for building a high-powered engineering organization from the ground up; including technical strategy, engineering infrastructure, customer engagement, engineering team building, processes development, and product road mapping. Prior to SOLUTE, Michael worked in the commercial software industry where he was responsible for engineering leadership at multiple successful startups and brings the startup mindset into the defense domain.
