The National Geospatial Intelligence Agency picked a systems integrator for a computer vision project. The integrator, CACI, started to develop a proprietary...
The National Geospatial Intelligence Agency picked a systems integrator for a computer vision project. The integrator, Consolidated Analysis Center, Incorporated (CACI), started to develop a proprietary function that a commercial company already offered. You can guess what happened. Everyone ended up in court. For the details on a case underscoring the legal preference for commercial products, Federal Drive with Tom Temin spoke with Haynes Boone procurement attorney Zach Prince.
Interview transcript:
Tom Temin All right, Zach, walk us through this, because it has to do with market research, it has to do with commercial preference and all of these kind of long standing issues.
Zach Prince Sure, Tom, it is complicated. So by statute, Congress has established a firm preference for commercial products and services. Agency heads are required when structuring procurements to set requirements so that they can have commercial products and services solutions, or at least non developmental solutions. And it goes further to require the agencies to instruct primes going further down the supply chain to incorporate commercial products and services as frequently as possible. In this case, NGA failed to do that.
Tom Temin What was NGA trying to acquire here?
Zach Prince So the National Geospatial Intelligence Agency, or the NAPS agency is a lot of folks still known as, obtains and analyzes a huge volume of images and other geospatial information. And then they use analytical tools to process that volume of data in order to obtain useful intelligence. And to do that, they use a system they call Computer Vision, which is a form of AI that trains and uses computers to interpret the visual world. So in this case, NGA was issuing a solicitation for an IDIQ contract reference, a sapphire. It had two components. It had a huge data repository to store and disseminate the geospatial intelligence, and a solution to integrate a computer vision system to enhance the agency’s ability to produce, review and classify the intelligence from all of these images.
Tom Temin So they needed the application and the storage as one deal.
Zach Prince That’s right. And they didn’t want to bifurcate this into storage and the actual system.
Tom Temin So they chose CACI, which proved it could do both, correct?
Zach Prince That’s right. CACI was able to do both. Technically, the solicitation allowed for either development or provision of an existing vision system. CACI chose the development option.
Tom Temin All right, so that sounds like custom development. Then there was this company called Percipient, which said, well, it could do the image capture part of it and the image processing part of it, but not the storage. So how did they get to be a party to this whole acquisition snafu?
Zach Prince So Percipient has a computer vision software called Mirage. It’s an open architecture software that, in their view, would fulfill NGA’s needs, which they told NGA about very early on. NGA told them to go talk to CACI, who at first said, sorry, it’s too late. NGA said, wait, wait, maybe it’s not too late. We’re still evaluating whether we even need a development solution. Maybe we’re just going to use our existing computer vision system. But ultimately, a Percipient had to hear from CACI at a trade show, that CACI had already been developing this new computer vision program. So NGA tried to stall them from filing a protest, told them to lay off the litigation and we’ll think about your software, but then told CACI to go ahead anyway and develop something that would do functionally the same thing.
Tom Temin That in itself doesn’t sound quite kosher to say, drop your lawsuit and we’ll consider buying from you. Can an agency actually say that? They said it, but are they allowed to say it, I guess I meant.
Zach Prince The court suggests that, No. But the decision here is on a motion to dismiss. So we don’t have anything on the merits yet.
Tom Temin Let’s back up a step. So Percipient went to the court of federal claims after this incident.
Zach Prince That’s right. So NGA did some evaluation of Mirage, but it was clear to Percipient that they weren’t taking it seriously. They were evaluating it as a different type of software than it actually is. And from Percipients logs of the use of Mirage, the agency really hadn’t gone in and done anything with it. So once it became clear that the agency was not going to be doing anything on its own to buy the software, they brought this protest.
Tom Temin All right. We’re speaking with Haynes Boone procurement attorney Zach Prince. And what did the court say and decide.
Zach Prince So as soon as the protest was filed, the agency and CACI moved to dismiss for lack of jurisdiction and for lack of standing and on timeliness grounds. Because remember, this is two years after the agency originally awarded the IDIQ to CACI, quite a while. And the court found that there was standing and there was jurisdiction, because of the commercial item preference that’s contained statute. And this is really following on the heels of the Federal Circuit’s Palantir decision from 2018, where they reprimanded the Army for failing to conduct market research into a commercial product that would have satisfied their needs. Instead, the Army in that case went out and procured a developmental software.
Tom Temin So in awarding the original contract to CACI, they weren’t really buying the total solution that they had put in the solicitation either, because CACI had to start coding this application, which might have existed commercially to begin with.
Zach Prince So the solicitation permitted that as a solution. Which is why the court said, actually it wouldn’t have been timely for Percipient to protest the solicitation, because the solicitation could have allowed for a commercial product solution. And so there are some very complicated and interesting jurisdictional and timeliness issues here, which I’m not going to go too far down the law nerd angle, but it is a case really worth watching, because commercial product providers frequently want their solutions to be part of a government procurement. They don’t want a development solution, but they also frequently can’t provide the full gamut of services an agency needs. But that isn’t enough to exclude that. The court hear, certainly, considered the possibility that the agency might be required to direct CACI to include Mirage as the solution for the computer vision.
Tom Temin But it hasn’t made that decision yet.
Zach Prince No, it hasn’t. So the briefings on the merits are going to take some time. I would strongly suspect there’ll be an appeal. So this will be something to watch for a while yet.
Tom Temin And I guess, maybe the other lesson here so far is that agencies should not take the requirement to do market research too lightly, especially in a burgeoning field with artificial intelligence and all kinds of imaging applications being developed. And the NGA itself is a agency always professing to want to use commercial imagery and commercial solutions. That’s kind of one of their mantras. And so that early piece of the market research cannot be ignored or taken lightly.
Zach Prince That’s right. And I think a lot of agencies before 2018 were taking it fairly lightly, because it didn’t seem like there was any legal teeth to the commercial product preference. But the federal circuit was clear that, in fact, that statute creates an enforceable right. And the court here is just continuing on the heels of that decision.
Tom Temin And there’s also the issue that CACI is a known quantity. It’s been around many, many, many years and has performed successfully for lots of agencies. It’s not like they’re a newcomer here. So that mitigates in favor of why the NGA would want to play it safe, maybe.
Zach Prince Maybe. But any time you’re developing a software solution, it’s going to be expensive and have risks. And certainly, the agency knows that. It might be the case that Mirage doesn’t do what the agency needs, but they still have to do that evaluation, which they still have failed to do.
Tom Temin And also, the company did not protest to the Armed Services Board of Contract Appeals, but rather to the Court of Federal Claims. Any significance there?
Zach Prince So it couldn’t have gone to the Armed Services Board, because it’s a bid protest. They might have gone to [Government Accountability Office (GAO)], but because of GAO’s strict timeliness rules, there’s no question that this would have been on timely at GAO.
Tom Temin Well, they had lawyers that knew what they were doing anyway, in terms of the venue for a protest. That’s sometimes as crucial as the merits, isn’t it?
Zach Prince It is. And the timeliness issues at the court are going to get very complicated, particularly in the heels of this decision. Because the court here, it’s commonly understood that at GAO you have these very strict five -10 day rules for when you can bring a protest. At the court, there’s this equitable thing called latches that you can’t bring a stale claim. But what that means is always been a little bit vague, but it’s allowed flexibility. The court here said actually the concept of latches is inapplicable in a bid protest based on a recent Supreme Court decision. I don’t know about that holding, the federal circuit seems to have already rejected that holding in a case last year. So we’ll see where that goes. But that would totally upset some of the basic rules about bid protest.
Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.
Tom Temin is host of the Federal Drive and has been providing insight on federal technology and management issues for more than 30 years.
Follow @tteminWFED