A Vision For the Future The initial successes and chal- lenges experienced with NIPRGPT illuminated a path toward a future where AI can empower SSC opera- tions with unprecedented speed, ef- ficiency, and strategic foresight. This vision extends beyond simply auto- mating individual tasks; it encom- passes a fundamental shift in how the USSF acquires, develops, and deploys space capabilities. It also addresses a critical underlying issue: the lack of standardized processes within many areas of space acquisition. Leveraging NIPRGPT, a defined workflow was created, complete with supporting artifacts like Excel templates and tailored instructions for different stakeholder groups. This demonstrates how AI can not only automate existing tasks but also fa- cilitate the creation of standardized processes where they are lacking, bringing much-needed structure and efficiency to complex projects. Be- yond process development, AI can transform core acquisition functions. Imagine AI-powered tools that op- timize satellite design parameters based on mission requirements and real-time data, automate launch
lenge was simply awareness and ac- cess. Many colleagues were unaware of NIPRGPT or how to access it for work. Furthermore, security restric- tions on the Air Force network pre- vented access to public LLMs like ChatGPT, limiting exposure to broader AI tools. This lack of awareness and access required proactive communi- cation and training. In the intervening months since these initiatives were taken, many more tools have come online as well as campaigns to in- crease workforce awareness. The security imperative. One of the greater advantages of NIPRGPT is its inherent security. Residing on government networks and developed by the government, it offers a level of data protection and confidential- ity that public LLMs simply cannot match. This is of paramount impor- tance within USSF and the DoW, where safeguarding sensitive infor- mation is crucial. Furthermore, hav- ing access to the source code allows for customization and deployment on secure internal systems, optimizing NIPRGPT for specific-use cases, and mitigating the risks associated with external platforms. The debugging dilemma. Integrat- ing NIPRGPT with existing workflows, particularly within Jira, presented a significant technical challenge. While NIPRGPT could generate code for au- tomating tasks like ticket creation, the process was far from seamless. The generated code often required exten- sive debugging due to inconsisten- cies between NIPRGPT’s output and the intricacies of Jira’s import func- tion, which was further complicated by Jira’s highly customizable nature and frequent updates. This debug- ging process, often using simple tools like WordPad, could be surprisingly time-consuming, sometimes negat- ing the time-saving benefits of using the LLM in the first place, especially when striving for perfect code. This highlighted the need for tighter inte- gration and more robust automation tools to fully realize the potential of AI-driven workflows.
scheduling by analyzing orbital dy- namics and resource availability, and streamline supply chain management by predicting and mitigating potential disruptions. This future is not about replacing humans with machines but about fos- tering a powerful synergy. The Infor- mation Operations Networking (ION) program provides a compelling exam- ple. A key challenge was the difficulty in staffing and training Model-Based Systems Engineering modelers. An innovative solution emerged: funding an AI-driven modeling effort. This ap- proach not only augments the capa- bilities of existing modelers but cre- ates opportunities for reskilling and retraining. Modelers can transition into AI model managers, overseeing and directing the AI while contributing their domain expertise. This not only increases efficiency but also allows those with modeling skills to contrib- ute to other critical areas within the organization, maximizing utilization of valuable talent. This exemplifies how human-machine teaming can enhance productivity and create new, more fulfilling roles within SSC. SSC recognizes the importance of responsible AI development and
Beyond process development, AI can transform core acquisition functions. Imagine AI-powered tools that optimize satellite design parameters based on mission requirements and real-time data, automate launch scheduling by analyzing orbital dynamics and resource availability, and streamline supply chain management by predicting and mitigating potential disruptions.
November-December 2025 | DEFENSE ACQUISITION | 21
Made with FlippingBook - Online Brochure Maker