icon-carat-right menu search cmu-wordmark
Our Research

Innovative Acquisition

The Department of Defense (DoD) needs more innovative acquisition strategies to deliver software capabilities with a speed that matches dynamic mission needs.

Defense acquisition is the set of processes the U.S. Department of Defense uses to obtain the systems and services necessary to support defense mission needs. CMU SEI is revolutionizing defense acquisition for software and systems—making it faster, leaner, and more innovative. Our work integrates data-driven insights, software engineering research, and acquisition science to rapidly create and deploy systems for warfighters and

  • help scale the software acquisition workforce
  • reduce the complexity of acquisition rules
  • enable decisive action
  • accelerate the adoption of commercial and scientific innovations

These CMU SEI capabilities ensure the Department of Defense provides warfighters with the mission capabilities needed to support the software-defined warfare environment. Our work promotes DoD–industry collaboration, fostering stronger partnerships to ensure mission success.

The SEI and the DoD Software Acquisition Pathway

Recognizing the imperative to modernize software acquisition policy, the Office of the Undersecretary of Defense for Acquisition and Sustainment (A&S) initiated work with the SEI. The SEI’s unique integration of data-driven insights, software engineering research, and acquisition science were critical to the October 2020 issuance of Department of Defense Instruction 5000.87 (DoDI 5000.87), the DoD’s Software Acquisition Pathway (SWP). The SWP departs from decades of hardware-based acquisition regulations. It helps the DoD acquire software by applying modern software practices, including Agile and DevSecOps, to deliver software capabilities with a speed that matches dynamic mission needs.

The SEI began working with the DoD on these challenging tasks in 2019 with a foundational Defense Innovation Board study and work on the design and execution of the FY18 National Defense Authorization Act (NDAA) Section 873 and 874 Agile pilot programs. The empirical results from the pilots informed the direction of SWP development, in accordance with Section 800 of the FY20 NDAA. The SEI team then worked hand-in-hand with stakeholders across the DoD and the defense industrial base in the iterative development, testing, and updating of the policy and supporting tools and resources.

Since the publication of the SWP in late 2020, the SEI has worked extensively with SWP adopters to effectively implement the SWP in different program contexts, identify barriers and challenges, and monitor outcomes. These activities create an important feedback loop that informs future policy refinement and the development of improved guidance and resources necessary to further accelerate and scale SWP adoption.

AI for Acquisition

AI augmentation is transforming software engineering, but there has been little focus on the impact of these transformative changes on software acquisition and the regulatory activities that are critical roles in DoD software ecosystems. The SEI is expanding the application of generative AI to accelerate and scale software acquisition. Stakeholders in software acquisition—who operate with challenging constraints of regulation, security, and accuracy—can benefit extensively from the use of generative AI, such as large language models (LLMs), if they understand how to adapt them to their workflow.

Innovation in software acquisition using generative AI is about increasing productivity for acquirers and stakeholders while mitigating risks. Through research and experiments, the SEI is helping DoD programs and acquisition professionals consider why, when, and how to start using these new tools. For example, LLMs excel in processing text and data, making them suitable for context-specific summarization and complex inquiries. This scalability is valuable when dealing with extensive software documentation, requirements, or codebases common in DoD acquisition programs.

Other benefits LLMs provide relate to their accessibility and ease of use. LLMs can be accessed through web services, APIs, and other platforms, making them readily available to acquisition programs. The ease with which they can integrate into existing workflows helps simplify their adoption in software acquisition. LLMs are also accessible to individuals with varying backgrounds using a natural language interface. This enables a wide range of nontechnical stakeholders to participate effectively in software acquisition.

Since LLMs are relatively new and fast-changing technology, the SEI’s work with LLMs and acquisition is essential for DoD programs and acquisition professionals to evaluate how LLMs may (or may not) align with their specific software acquisition needs critically and objectively, as well as formulate strategies to address potential risks.

What We Offer

The Latest from the SEI Blog

Perspectives on Generative AI in Software Engineering and Acquisition

Blog Page
, , , , , and

It is essential that software and acquisition professionals learn how to apply AI-augmented methods and tools in their workflows. SEI researchers offer their perspectives on this topic.

READ

The Myth of Machine Learning Non-Reproducibility and Randomness for Acquisitions and Testing, Evaluation, Verification, and Validation

Blog Page
, , , , and

A reproducibility challenge faces machine learning (ML) systems today. This post explores configurations that increase reproducibility and provides recommendations for these challenges.

READ

Latest from the Digital Library

DoD Software Acquisition Pathway: Modernizing and Speeding Software Acquisition

Fact Sheet
Software Engineering Institute

CMU SEI research has focused on driving changes to help scale the software acquisition workforce, reduce the complexity of acquisition rules, and adapt acquisition practices to enable decisive action, accelerating the adoption of commercial and scientific innovations.

Learn More

Capability-Based Software Cost Estimation (CaBSCE): Modernizing Software Cost Estimates for Agile and DevSecOps (Fact Sheet)

Fact Sheet
Software Engineering Institute

The SEI is developing the Capability-Based Software Cost Estimation method to modernize cost estimation practices by aligning them with Agile and DevSecOps.

Learn More