Please use this identifier to cite or link to this item: https://dair.nps.edu/handle/123456789/5528
Title: Augmented Acquisition: Operationalizing Human-in-the-Loop AI to Accelerate Warfighting Capabilities
Authors: Omar Haroun, Ashish Agrawal
Dylan Mrla
Keywords: Defense acquisition
human-centered AI
augmented intelligence
contracting officer decision support
Procurement Acquisition Lead Time
human-in-the-loop AI
39th Contracting Squadron
acquisition acceleration
Issue Date: 30-Apr-2026
Publisher: Acquisition Research Program
Citation: APA 7
Series/Report no.: Acquisition Management;SYM-AM-26-091
Acquisition Management;SYM-AM-26-212
Abstract: Contracting officers are the backbone of delivering mission-critical advantages to the warfighter, responsible for navigating thousands of pages of regulatory guidance across the Federal Acquisition Regulation (FAR) and its sub-regulatory supplements while producing accurate, defensible, and timely acquisition packages. Beyond regulatory complexity, contracting professionals must manage massive volumes of contract data across the enterprise and at the squadron level while often simultaneously operating across multiple disconnected systems to accomplish a single task. The acquisition process itself spans numerous stages, from requirements development through pre-solicitation, solicitation, pre-award, and award, each with distinct documentation and compliance demands. Compounding these challenges, requirement owners frequently delay timelines of critical acquisitions due to poorly defined or incomplete requirements, a key area where AI can provide meaningful augmentation. As the Department of War accelerates efforts to deliver warfighting capability at the speed of relevance, the acquisition workforce faces mounting pressure to reduce Procurement Acquisition Lead Time (PALT) without sacrificing compliance or accountability. The leading approach to address these challenges with cutting-edge AI technology is to combine human input with artificial intelligence to significantly enhance the productivity of high-performing personnel. This paper investigates an alternative paradigm: human-centered AI that augments contracting officer judgment rather than replacing it. The proposed approach positions AI as a copilot that reviews, validates, and enhances human-generated content in real time, enabling acquisition professionals to reach their maximum capacity for production while retaining full ownership of every decision. Our research is grounded in an active research and development (R&D) contract with the 39th Contracting Squadron (39 CONS) at Incirlik Air Base, where Eudia is developing and evaluating an augmented acquisition capability comprising three integrated applications: Insights, Sigma, and Augmented Contract Review (ACR). The system reviews human-generated acquisition documents against structured training data drawn from the FAR, DFARS, DAFFARS, internal policy, and historical contract files, providing clear citations, plain language explanations of regulatory risks, and structured feedback that strengthens document quality and compliance confidence. Through testing, we estimate PALT reductions varying by acquisition type—with simpler actions such as Simplified Acquisition Procedures under $250,000 showing different improvement rates than complex Source Selection efforts. Preliminary findings indicate that human-in-the-loop AI reduces document error rates, accelerates regulatory review, and supports workforce development by reinforcing training concepts and surfacing relevant guidance for junior personnel. This research contributes to the symposium’s focus on accelerating warfighting capability by demonstrating that the fastest path to contract award is not replacing human judgment but amplifying it.
Description: Presentation and Excerpt
URI: https://dair.nps.edu/handle/123456789/5528
Appears in Collections:Annual Acquisition Research Symposium Proceedings & Presentations

Files in This Item:
File Description SizeFormat 
SYM-AM-26-091.pdfExcerpt814.09 kBAdobe PDFView/Open
SYM-AM-26-212.pdfPresentation1.9 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.