percipient.ai, Inc. v. United States

CourtUnited States Court of Federal Claims
DecidedApril 7, 2023
Docket23-28
StatusPublished

This text of percipient.ai, Inc. v. United States (percipient.ai, Inc. v. United States) is published on Counsel Stack Legal Research, covering United States Court of Federal Claims primary law. Counsel Stack provides free access to over 12 million legal documents including statutes, case law, regulations, and constitutions.

Bluebook
percipient.ai, Inc. v. United States, (uscfc 2023).

Opinion

In the United States Court of Federal Claims No. 23-28C (Filed: March 31, 2023) (Re-filed: April 7, 2023) 1 ************************** PERCIPIENT.AI, INC.,

Plaintiff,

v.

THE UNITED STATES,

Defendant,

and

CACI, INC. – FEDERAL,

Intervenor. **************************

Samuel C. Kaplan, Washington, DC, for plaintiff, Percipient.ai, with whom were Hamish P.M. Hume, Eric J. Maurer, and Gina A. Rossman, of counsel.

Reta E. Bezak, Senior Trial Counsel, United States Department of Justice, Commercial Litigation Branch, with whom were Brian M. Boynton, Principal Deputy Assistant Attorney General, Patricia M. McCarthy, Director, and Corinne A. Niosi, Assistant Director, for defendant. Graham Day, National Geospatial-Intelligence Agency, of counsel.

Anne B. Perry, Washington, DC, for intervenor, CACI, Inc. – Federal, with whom was Jonathan S. Aronie and Ariel E. Debin, of counsel.

1 This opinion was originally issued under seal to give the parties an opportunity to propose redactions. Because the parties agreed that none were necessary, the opinion appears in full. OPINION This is a post-award bid protest of the National Geospatial- Intelligence Agency’s alleged violation of 10 U.S.C. § 3453, a statute that requires agencies to procure commercial or non-developmental products “to the maximum extent practicable.” Both the United States and the intervenor, CACI, Inc. – Federal, move to dismiss the protest for lack of subject-matter jurisdiction.

The matter is fully briefed, and oral argument was held on March 6, 2023. We denied the motions to dismiss in an order issued on March 9, 2023. This opinion more fully explains our reasoning.

BACKGROUND 2

The National Geospatial-Intelligence Agency (NGA) obtains and analyzes images and other geospatial information to provide the federal government with intelligence data. Supplying this kind of intelligence on a global scale is a burdensome analytical task and cannot be done effectively without the help of advanced computer technology. One of those advanced technologies is computer vision, a form of artificial intelligence that “trains and uses computers to interpret the visual world.” Compl. ¶ 55. With computer vision, users can more efficiently compile and analyze geospatial intelligence.

Hoping to benefit from this technology, NGA, more than three years ago, issued the SAFFIRE solicitation—which was an indefinite delivery, indefinite quantity contract containing two parts. The first was a data repository that would store and disseminate geospatial intelligence “across various large organizations.” Compl. ¶ 60. The second, which is at the heart of this dispute, would integrate a computer vision system to enhance the

2 When a party moves to dismiss for lack of subject matter jurisdiction, the court assumes that the undisputed facts in the complaint are true and draws reasonable inferences in the plaintiff’s favor. Acevedo v. United States, 824 F.3d 1365, 1368 (Fed. Cir. 2016). These undisputed facts are drawn from the complaint, the attached materials, and the administrative record. 2 agency’s ability to produce, review, and classify intelligence from “millions” of images. Compl. ¶ 58.

The plaintiff, Percipient, is a technology company that developed a computer vision software called “Mirage.” Mirage is an open architecture software that works alongside other computer systems and can detect equipment, vehicles, and faces—each of which is a critical aspect of geospatial intelligence. More than that, though, Mirage’s tools also allow users to narrow the computer’s focus to specific objects, patterns, or geographical areas, and it can even learn to anticipate its users’ needs over time. Despite these features and capabilities, Percipient did not bid on the SAFFIRE contract because its software could only fulfill SAFFIRE’s computer vision requirements, not the entire contract. For that reason, Percipient relied on what it viewed as the agency’s statutory obligation to consider incorporating commercial products and hoped to be part of NGA’s SAFFIRE efforts.

In January 2021, NGA awarded the SAFFIRE contract to CACI and informed Percipient that if it wanted to participate in SAFFIRE, it needed to speak with CACI. This eventually led to a meeting between Percipient and CACI in March 2021. At this meeting, CACI expressed significant interest in partnering with Percipient on future projects, but explained that, as for working together on SAFFIRE, “that ship” had already “sailed.” Compl. ¶ 93.

Alarmed by this revelation, Percipient asked NGA if it would independently evaluate Mirage as a possible commercial solution for SAFFIRE’s computer vision system. NGA responded several weeks later and reassured Percipient of its commitment to using commercial products. NGA further explained that CACI’s “ship has sailed” statement was an “unfortunate miscommunication” that did not reflect the agency’s position. Compl. ¶ 100. Instead, the agency had not yet decided whether it needed to incorporate a commercial product because CACI was still reviewing NGA’s legacy systems. NGA confirmed that commercial products would be evaluated once CACI finished.

Another two months went by before Percipient finally secured a meeting with CACI to demonstrate Mirage, although CACI’s Program Manager—the individual largely responsible for deciding whether to 3 incorporate a commercial product—left the meeting after only 20 minutes. Still, Mirage received positive feedback, and CACI promised to evaluate Mirage more fully. This “deep dive” into Mirage never happened, however. Compl. ¶ 108.

Several months later, Percipient learned at the 2021 GEOINT Symposium that CACI would be developing a computer vision system for SAFFIRE when CACI employees visited Percipient’s symposium booth. Surprised by the news, and no longer believing that CACI could fairly evaluate Mirage, Percipient met with NGA and asked to set up a demonstration. NGA agreed but requested that Percipient “ease up on the legal pressure.” Compl. ¶ 118. Percipient then demonstrated Mirage’s abilities to several NGA representatives in December 2021, at the end of which NGA concluded that Percipient’s software met “all of NGA’s analytical transformation requirements.” Compl. ¶ 120.

Over the next several months, the parties worked to reach an agreement that would allow NGA to test Mirage with live data, something that Percipient agreed to do at no cost. Just before signing an agreement to that effect, however, NGA changed its tune. Citing legal and security complexities, NGA would no longer use live data and would instead use previously released and publicly available images. Percipient pushed back, claiming that these images would not allow NGA to test Mirage’s geospatial module or some of its unique features, like its ability to alert changes over time. After significant delay, NGA relented and allowed the use of live data.

NGA completed its testing of Mirage in October 2022. Based on the results, Percipient suspected that NGA was not assessing Mirage as a possible commercial solution for SAFFIRE’s computer vision requirements because, among other reasons, Percipient could only identify four NGA searches over the 12-week testing period. Thus, Percipient offered to extend the testing period, again at no cost, so NGA could more fully evaluate Mirage as a computer vision system. Percipient’s suspicions appeared to be confirmed, though, when NGA explained one month later that it had evaluated Mirage as “an enterprise Machine Learning Platform,” and not “as an Analytical tool.” Compl. ¶ 137.

Free access — add to your briefcase to read the full text and ask questions with AI

Related

United States v. MacK
295 U.S. 480 (Supreme Court, 1935)
Baker v. Carr
369 U.S. 186 (Supreme Court, 1962)
Richlin Security Service Co. v. Chertoff
553 U.S. 571 (Supreme Court, 2008)
Carcieri v. Salazar
555 U.S. 379 (Supreme Court, 2009)
Weeks Marine, Inc. v. United States
575 F.3d 1352 (Federal Circuit, 2009)
Distributed Solutions, Inc. v. United States
539 F.3d 1340 (Federal Circuit, 2008)
Blue & Gold Fleet, L.P. v. United States
492 F.3d 1308 (Federal Circuit, 2007)
Ramcor Services Group, Inc. v. United States
185 F.3d 1286 (Federal Circuit, 1999)
American Fed. of Govt. Employees v. United States
258 F.3d 1294 (Federal Circuit, 2001)
Anderson v. United States
344 F.3d 1343 (Federal Circuit, 2003)
McAfee, Inc. v. United States
111 Fed. Cl. 696 (Federal Claims, 2013)
Petrella v. Metro-Goldwyn-Mayer, Inc.
134 S. Ct. 1962 (Supreme Court, 2014)
Federal Acquisition Services Team, LLC v. United States
124 Fed. Cl. 690 (Federal Claims, 2016)
Acevedo v. United States
824 F.3d 1365 (Federal Circuit, 2016)
Palantir Usg, Inc. v. United States
904 F.3d 980 (Federal Circuit, 2018)
Sekri, Inc. v. United States
34 F.4th 1063 (Federal Circuit, 2022)

Cite This Page — Counsel Stack

Bluebook (online)
percipient.ai, Inc. v. United States, Counsel Stack Legal Research, https://law.counselstack.com/opinion/percipientai-inc-v-united-states-uscfc-2023.