The U.S. Army is continuing to pursue changes in the way it approaches software development.
During an AFCEA NOVA online seminar on May 9, Army Chief Information Officer Leo Garciga outlined how a recently established policy is driving the adoption of agile software development practices.
In early March, Army leadership released a new policy that is enabling the adoption of modern software development and acquisition practices. The new policy institutionalizes modern software development approaches across the service branch, which focuses on iterative development and delivery of software.
“We’re working to make sure that we shape [software] in the right way so we can have that acceleration to get onto these platforms, as opposed to the more legacy mindset, which is, “hey, we’re going to do a lot of software development, and everything needs to be custom because we’re special,” Garciga said.
Software has gone from being a means of operating hardware-based weapons systems to being the essential driver of much of the Army’s weapons, business, and training platforms. Therefore, the Army needs to rapidly develop and upgrade software-based capabilities to maintain a competitive advantage.
Garciga explained that the new policy reforms many of the institutional processes that underpin the former software development lifecycle to accelerate the adoption of modern software development, including changing the way requirements for software systems are written, the types of acquisition strategies and buying models employed, and how the Army writes its requests for proposals and contracts with vendors.
“The policy is demanding that we write our need statements at a much higher level than we ever have,” Garciga said. “Theoretically, that could open the door to new vendors, new commercial solutions, and to higher level needs.”
Another key institutional process that the policy highlights is data management. According to Garciga, interoperable, interchangeable software cannot be developed “without the data layer being done correctly.”
“[Data] is foundational to being able to develop the proper software,” he said.
The new policy only applies to new contracts moving forward, the CIO said. This includes contracts in the negotiation phase, those not yet finalized, and those coming up for re-compete on a program.
“Our program offices are looking to determine those contract types, contract approaches, etc. [and determine] if the terms and conditions should be modified to better enable modern software development. So, I think it’ll vary a little bit across the enterprise, depending on where all of these different programs are,” Garciga said.
In other news, Garciga gave a sneak peek at a generative artificial intelligence (GenAI) and large language models (LLM) memo his office is working on that would help guide the adoption of these new technologies into the service.
According to Garciga, the upcoming memo will include guidance on data protection, AI guardrails, and industry collaboration.
“We’re going to start the initial wrapping of what that’s going to look like in our initial guidance memo, so we can start having a more robust conversation in this space,” he said. “It’s pretty beefy … and we hope to have that out very quickly.”
