Other-transaction authority, codified for the Defense Department in 10 U.S.C. § 4021 and related provisions, permits awards outside the Federal Acquisition Regulation when the work is a prototype or, under later amendments, a follow-on production effort that began as a prototype. Public budget justifications and GAO testimonies describe the vehicle as a response to commercial software vendors that declined FAR-based contracts. According to the Department’s own reports to Congress, OT obligations rose sharply in the late 2010s and early 2020s, with software, cloud, and data-platform efforts among the largest categories.

GAO’s acquisition work, including reports on prototype OTs and software pathway pilots, has repeatedly found that agencies did not always document why an OT was preferred to a FAR contract, how competition was conducted among more than one source, or how follow-on production awards were priced. The record shows GAO recommending stronger guidance on intellectual-property assertions, data-rights clauses, and the conditions under which a prototype OT may convert to a production OT without a new competition. Defense officials have publicly described OTs as necessary to match commercial development cycles; GAO has described the same awards as carrying residual risk when cost, schedule, and performance baselines are not established in a manner comparable to major defense acquisition programs.

Software-specific findings appear in GAO reviews of the Adaptive Acquisition Framework’s software pathway and of selected CDAO and service software-factory efforts. According to those reports, some awards bundled platform licenses, integration labor, and cloud consumption under a single OT, making it difficult for auditors to isolate unit costs or to determine whether follow-on production was limited to the prototype’s original scope. IG audits of selected OT consortia have alleged that membership fees and internal down-selects reduced the appearance of full and open competition even when multiple vendors participated in a consortium. The Department has stated in public comments on draft GAO reports that consortium models expand the industrial base; GAO has stated that the Department could not always demonstrate that non-consortium firms had a fair opportunity.

Budget exhibits for research, development, test, and evaluation and for operation and maintenance show line items for software factories, DevSecOps pipelines, and tactical cloud that are executed in part through OTs rather than through traditional systems-engineering contracts. Public dockets do not always map those line items to a named program of record with a milestone decision authority. GAO has noted that the absence of such mapping complicates congressional oversight of whether prototype spending later appears as procurement or as sustainment. Officials have said that software is never “done” and that OTs allow continuous iteration; GAO has said that continuous iteration does not relieve the Department of the requirement to show that prices paid remain reasonable over time.

Follow-on production OTs, authorized when a prototype was successfully completed and competition was used to the maximum extent practicable, have drawn particular GAO attention. Findings include instances in which the prototype phase was brief, the production phase large, and the documentation of “successful completion” limited to a letter from a program office. The record shows GAO recommending that the Department define success criteria in the prototype agreement itself. Defense acquisition officials have publicly accepted some recommendations and disagreed with others, arguing that commercial software terms cannot always be restated in government-unique language without losing vendors.

Intellectual-property terms remain a recurring point of tension. Public GAO summaries describe awards in which vendors retained rights in preexisting commercial code while the government received limited government-purpose rights in modifications. Auditors alleged that the Department sometimes lacked an inventory of what was preexisting versus newly developed, complicating later recompete or in-sourcing. The Department has said that commercial licenses are the price of access to modern tooling; GAO has said that without an inventory the government cannot assess lock-in.

Congressional committee reports accompanying recent National Defense Authorization Acts have directed additional reporting on OT use, including software, and have asked GAO to continue reviewing competition and follow-on practices. No public GAO product claims that OTs are unlawful; the findings concern implementation, documentation, and the risk that speed is purchased at the cost of later bargaining power. The Department’s published OT guide and subsequent updates incorporate some of those recommendations. Whether the volume of software OTs will decline as FAR-based software pathways mature is not established in the public record; budget justifications continue to list both vehicles.

The comparison that emerges from open sources is not a binary of success or failure. It is a documented expansion of a statutory exception, paired with a documented series of GAO and IG observations that the exception’s safeguards—competition to the maximum extent practicable, clear prototype-to-production criteria, and auditable price and IP terms—were applied unevenly across software awards. Further reporting, according to GAO’s own open recommendations list, remains due on whether those observations have produced measurable changes in award files.