C3.ai, Inc.

B-418676: Jul 28, 2020

Additional Materials:


Ralph O. White
(202) 512-8278

Kenneth E. Patton
(202) 512-8205


Office of Public Affairs
(202) 512-4800

C3.ai, Inc., of Redwood City, California, protests the terms of request for task order proposals (RFTOP) No. 832015495, which was issued by the Defense Information Systems Agency (DISA), Joint Artificial Intelligence Center (JAIC), for the creation and operation of a cloud-based platform where artificial intelligence and machine learning (AI/ML) software and data will be integrated and housed for the Department of Defense. The protester asserts that DISA should obtain the products and services as a commercial item acquisition under Federal Acquisition Regulation (FAR) part 12. The protester also contends the RFTOP is outside the scope of the underlying indefinite-delivery, indefinite-quantity (IDIQ) contract, and that the RFTOP impermissibly increases the maximum value of the underlying contract.

We deny the protest.

The decision issued on the date below was subject to a GAO Protective Order. This redacted version has been approved for public release.


Matter of:  C3.ai, Inc.

File:  B-418676

Date:  July 28, 2020

Joseph D. West, Esq., Lindsay M. Paulin, Esq., and Crystal L. Weeks, Esq., Gibson, Dunn & Crutcher LLP, for the protester.
Colleen A. Eagan, Esq., and Vera A. Strebel, Esq., Defense Information Systems Agency, for the agency.
Stephanie B. Magnell, Esq., and Evan C. Williams, Esq., Office of the General Counsel, GAO, participated in the preparation of the decision.


1.  Protest that the agency improperly consolidated commercial and non-commercial items in a single task order procurement, and should instead have obtained the commercial items in separate procurements using the commercial items procedures of Federal Acquisition Regulation part 12, is denied, where the agency had a reasonable basis for the selected procurement structure. 

2.  Protest that task order proposal request exceeds the scope of the underlying multiple-award indefinite-delivery, indefinite-quantity contract is denied where the record shows the services sought were reasonably encompassed within the contract’s scope of work.


C3.ai, Inc., of Redwood City, California, protests the terms of request for task order proposals (RFTOP) No. 832015495,[1] which was issued by the Defense Information Systems Agency (DISA), Joint Artificial Intelligence Center (JAIC), for the creation and operation of a cloud-based platform where artificial intelligence and machine learning (AI/ML) software and data will be integrated and housed for the Department of Defense.  The protester asserts that DISA should obtain the products and services as a commercial item acquisition under Federal Acquisition Regulation (FAR) part 12.[2]  The protester also contends the RFTOP is outside the scope of the underlying indefinite-delivery, indefinite-quantity (IDIQ) contract, and that the RFTOP impermissibly increases the maximum value of the underlying contract.

We deny the protest. 


Underlying IDIQ Contract

In 2018, DISA concluded its procurement for the Systems, Engineering, Technology and Innovation (SETI) IDIQ contract, making award to 15 large businesses.[3]  COS/MOL at 10.  The SETI contract has eight task areas:  (1) systems engineering; (2) design analysis engineering; (3) systems architecture; (4) software systems design and development; (5) systems integration; (6) systems test and evaluation; (7) systems deployment and life-cycle engineering; and (8) special systems engineering requirements.  SETI Contract at 21.  The total period of performance of the SETI contract is 10 years, consisting of one 5-year base period and one 5‑year option period.  Id. at 3-6.

The maximum value of the SETI contract is $7.5 billion, divided into contract line item numbers (CLINs) with different types of payment structures, such as fixed-price, and cost-reimbursement.  Id. at 2-10.  As relevant here, under the SETI contract the total value of all cost-reimbursement CLINs was capped at $75 million per contract period.  Id. at 3, 6.  At award, the total contract value was distributed equally among awardees, but the agency was permitted to later “reallocate the capacity amongst the contract holders, if necessary.”  Id. at 2. 

Contested Task Order

On November 22, 2019, the agency finalized its JAIC Mission Product Requirements, with the goal of obtaining “a prime product and services integrator to carry the bulk of coordinating activities . . . .”  AR, Tab 4, JAIC Mission Product Requirement at 18.  Also on November 22, DISA posted on the System for Award Management,a sources sought notice for the procurement that would become the JCF solicitation.  COS/MOL at 7.  The notice sought responses from entities “who can provide System Engineering and Third Party System Integration services to support the procurement, implementation and operation of a hybrid and multi-cloud deployable development and production platform for [AI/ML] solutions and applications across the Department of Defense [DOD] at enterprise scale in all security enclaves.”  AR, Tab 5a, Sources Sought Notice at 1; see also Tab 10, April Market Research Report at 4.  The JCF was “envisioned to be a cloud based AI/ML development platform where Data, Tools and Processes for the [DOD]/Government shall be encapsulated.”  AR, Tab 5a, Sources Sought Notice at 1.  Interested firms were to describe their experience in:

1)  “building out cloud infrastructure for enterprise scale and connecting it to external data repositories in a DevSecOps [development, security, and operations] framework”;

2)  “utilizing commercial cloud services providers and heterogeneous [DOD] assets e.g. high performance computing systems and edge systems”;

3)  “user management”;

4)  “guiding a program through the . . . process for cybersecurity compliance . . . toward achieving an Authority To Operate” (ATO);[4]

5)  “meeting the cybersecurity compliance requirements across . . . security enclaves”;

6)  “continual engineering support . . . to end users, system owners, system developers and help desk personnel”; [and]

7)“configuring or testing the capabilities of open source and commercial tools[.]”

AR, Tab 5A, Sources Sought Notice at 1; see also Tab 10, April Market Research Report, Critical Performance Requirements at 2-3. 

The agency received responses from 24 small businesses and 24 large businesses, including C3.ai, by the December 20 response date.  AR, Tab 10, April Market Research Report at 4; Tab 6j, C3.ai Sources Sought Response.  The agency concluded that only six firms, none of which was C3.ai, satisfied the agency’s critical performance requirements.[5]  Id. at 4-7.  With respect to C3.ai, of the seven topic areas above, the agency determined that the protester did not demonstrate experience in two areas:  [DELETED] and [DELETED].  AR, Tab 10, April Market Research Report at 5.  In addition, at the time of the market research, C3.ai did not demonstrate that it was likely to have [DELETED] by the time of award.[6]  Id.; see also Tab 5a, Sources Sought Notice, Special Requirements at 3. 

As relevant to this protest, on February 27, 2020, C3.ai announced that it was partnering with IBM and, “[u]nder the terms of the agreement, IBM will become the first preferred global systems integrator for the C3.ai platform.”  AR, Tab 13, C3.ai Press Release at 2.

On March 3, the agency determined that the procurement was commercial because “all services under this requirement are available to the general public and are offered through commercial companies.”  AR, Tab 25, March Market Research Report at 10.  This market research report was signed by a contracting officer temporarily assigned to the project, because the assigned contracting officer was completing another project.  Tr. at 102. 

On March 9, after reviewing a draft of the RFTOP and the market research report, the SETI contract program manager determined that the JCF requirement was in the scope of the SETI contracts.  AR, Tab 9, SETI Scope Determination.

On March 12, the assigned contracting officer reviewed the responses to the sources sought notice and the results of outreach to the Defense Innovation Unit.[7]  AR, Tab 11, Non-Commercial Memorandum for Record at 2.  The contracting officer noted that, while many AI/ML platforms were commercial, there also was a need for the contractor to “enabl[e] AI platforms to scale across all [DOD] missions and integrat[e] AI platforms into an authorization to operate environment for handling [DOD] controlled and sensitive data.”[8]  Id. at 1.  Accordingly, the contracting officer identified a need for “a JCF prime integrator with the responsibility of bringing in commercial solutions for rapidly authorizing AI platforms and tools.”  Id. at 3.  The contracting officer concluded that the procurement was non-commercial because “[n]o singular commercial solution addresses all capability areas. . . .”[9]  The contracting officer selected the SETI contract as the procurement vehicle because it was held by four of six capable respondents and there was “immediate cost-savings . . . in that there is no contract access fee[.]”  Id.

On March 13, after reviewing the market research, the agency affirmed its conclusion that the RFTOP was “non-commercial in nature.”  AR, Tab 12, Acquisition Plan at 5.  In selecting a cost-plus-fixed-fee contract type, DISA further noted that it was “not possible to estimate accurately the extent or duration of the time required to perform the work for all requirements under this contract at the time of contract award.”  Id. at 6; Tab 12a, Acquisition Plan Signature Sheet.   

On March 25, DISA issued the RFTOP under the procedures of FAR section 16.505 to all holders of the SETI large business contract.  The estimated value of the procurement is approximately $[DELETED] million.  AR, Tab 10, Market Research Report at 1 (estimating $[DELETED] million); Tab 7, Independent Government Cost Estimate (IGCE), Tab 2 (estimating $[DELETED] million).  The RFTOP, with a period of performance of one 12-month base period and three 12-month option periods, will compensate the contractor on cost-plus-fixed-fee and cost-reimbursement bases.  RFTOP at 2.[10]  Award is to be made to the offeror whose proposal is determined to represent the best value to the agency using a multi-stage best-value tradeoff evaluation process considering past performance, technical/management approach, and cost.  Id. at 4-6.  The past performance factor is a threshold factor, evaluated on an acceptable/unacceptable basis.  Id. at 4-5.  Offerors receiving an acceptable rating for past performance would be invited to provide an oral presentation and submit a written technical/management and cost proposal.  Id.  The best-value tradeoff decision would be based on the technical/management approach and cost factors, with technical/management approach more important than cost.  Id. at 6. 

On April 2, the JCF branch chief for strategy and development, a computer engineer who serves in that role and as the JCF chief of architecture and the JCF acquisition leader, and the contracting officer addressed the documentation of the non-commercial nature of the procurement.  The JCF branch chief wrote that he “misunderstood the determination” in the March market research report that the solicitation was commercial, and that there was no single vendor who could provide a single product that would satisfy the agency’s needs.  AR, Tab 27, Email from DISA Contracting Officer to JCF Branch Chief, Apr. 2, 2020, at 1; see also Tr. at 7.  He clarified that the contractor would integrate commercial capabilities, but the solicitation’s complexities and security requirements rendered it non-commercial.  Id. 

In approximately late March/early April, the agency revised the market research report to reflect that the services sought were “non-commercial due to the infancy of this type of cloud environment support in the [DOD].”[11]  AR, Tab 10, April Market Research Report at 10.  In this regard, the agency found as follows:

Although these services are commercial in nature, the tasks are required to be performed in a Government environment with unique security requirements.  This service is not commercially available in the [DOD] marketplace and therefore is managed and procured based on direct [DOD] requirements from qualified vendors.  Due to the clearance requirements and performance in the Government environment, the services are non-commercial.


Proposals addressing offerors’ technical/management approach and costs were due on April 22.  This protest was timely filed on April 20.  On June 18, the GAO attorney assigned to the protest held a telephonic hearing in order to obtain testimony from the JCF branch chief.[12] 


C3.ai first challenges the agency’s decision not to conduct the procurement using the procedures for procuring commercial items under FAR part 12.  In this regard, the protester contends that DISA failed to determine whether its needs could be satisfied by commercial solutions, as required by statute and regulation, improperly consolidated requirements in order to render the procurement non-commercial, and did not contemplate or allow for the proposal of commercial solutions in the JCF RFTOP.  As explained below, we find that, although the contemporaneous record is not entirely clear as to the order of events, and the documentation could have been more robust, overall, the protester has not shown that the agency’s conclusions were unreasonable.

The protester also asserts that the RFTOP is outside the scope of the underlying SETI contract’s technical requirements.  As addressed below, we conclude that the RFTOP tasks were reasonably within the scope of services described in the SETI contract.[13] 

Commercial Item

C3.ai contends that DISA should have met its requirements using commercial items procedures.  As background, the Federal Acquisition Streamlining Act of 1994 established, among other things, a preference and specific requirements for the acquisition of commercial items that are sufficient to meet the needs of an agency.  Federal Acquisition Streamlining Act of 1994 (FASA), Pub. L. No. 103-355 § 8104, 108 Stat. 3243 (codified, as amended, at 10 U.S.C. § 2377).  The FASA, as implemented by FAR part 12, requires agencies to conduct market research to determine whether commercial items exist that could meet the agency’s requirements; be modified to meet the agency’s requirements; or meet the agency’s requirement if the requirements were modified to a reasonable extent.  10 U.S.C. § 2377(c)(2);FAR 12.101.  Determining whether a product or service is a commercial item is largely within the discretion of the contracting agency, and such a determination will not be disturbed by our Office unless it is shown to be unreasonable.  Coulson Aviation (USA) Inc., B‑414566, July 12, 2017, 2017 CPD ¶ 242 at 4; Analytical Graphics, Inc., B‑413385, Oct. 17, 2016, 2016 CPD ¶ 293 at 12.

In reviewing the market research, DISA found that the commercial solutions submitted in response to the sources sought notice generally did not demonstrate experience in all of the subject matter areas of the notice.  AR, Tab 11, Memorandum for Record at 2.  In contrast, in almost all cases the companies that satisfied the critical performance requirements did so by proposing to use a partner.  The agency determined that no single commercial solutions existed to meet the totality of the requirements and decided to “establish a JCF prime integrator with the responsibility of bringing in commercial solutions for rapidly authorizing AI platforms and tools.”  Id. at 3.

Here, C3.ai, like many other firms, did not propose a single commercial item to satisfy the agency’s requirements.  See, e.g., Comments at 86 (asserting that “C3.ai can satisfy all of the [DELETED] PWS requirements, or [DELETED]% of the total PWS requirements[.]”).  While the protester makes a standalone commercially-available AI/ML platform, in order to meet all of the JCF requirements, it anticipated partnering with a firm that provides integrator services.  AR, Tab 6j, C3.ai Sources Sought Response at 2, 9; Tab 13, C3.ai Press Release, Feb. 27, 2020.  Accordingly, even if the agency were to attempt to solicit its requirements in a single consolidated commercial item procurement, C3.ai could not compete because it does not perform a critical part of those requirements. 

Our Office has consistently explained that to prevail, a protester must demonstrate that it has been competitively prejudiced by the agency’s errors.  Where the protester fails to demonstrate that, but for the agency’s actions, it would have had a substantial chance of receiving the award, there is no basis for finding prejudice, and our Office will not sustain the protest, even if deficiencies in the procurement are found.  Coulson Aviation, (USA), Inc., supra at 6.  Here, because C3.ai has not demonstrated that it offers a single commercial product that meets the agency’s requirements, we find no basis to conclude that the protester was prejudiced by the agency’s decision not to seek a single commercial item to satisfy its requirements.

Consolidation of Requirements

Next, C3.ai argues that the agency improperly consolidated commercial and non-commercial requirements in a manner that unduly restricts competition.  Protest at 6.  The protester then claims that if no single commercial solution exists to meet DISA’s needs then, under FASA and the Competition in Contracting Act of 1984 (CICA), “the Agency was obligated to consider whether it could modify its requirements – by, for example, unbundling them.”  Comments at 65-67.  The protester’s argument that the various facets of the procurement should be cleaved into their unique components is a necessary precursor to its claim that the agency use commercial item procurement procedures for the commercial components.  Protest at 6; see also Comments at 61.  The question here therefore centers on whether the agency had a reasonable basis to obtain its solution for commercial and non-commercial items in a single procurement.[14] 

DISA acknowledges that the RFTOP combines commercial and non-commercial requirements, but asserts that it has a reasonable basis for doing so.  COS/MOL at 35.  In particular, the agency asserts that there are particular functions for ATO approval and software interoperability that would not be satisfied if the agency tried to meet its requirements under separate procurements.  Id.

Our decisions provide that a contracting agency has the discretion to determine its needs and the best method to accommodate them.  General Electrodynamics Corp., B‑298698, B-298698.2, November 27, 2006, 2006 CPD ¶ 180 at 3.  This principle applies to the contracting format used to purchase the items which the agency has determined necessary.  Voith Hydro, Inc., B-401244.2, B-401771, Nov. 13, 2009, 2009 CPD ¶ 239 at 4.  Our Office will not object to an agency’s determination in this regard unless the protester shows that it is clearly unreasonable.  Id. 

When an agency seeks to procure separate and multiple requirements under a single contract, there is potential for restricting competition by excluding firms that furnish only a portion of the requirement; we therefore review challenges to such solicitations to determine whether the approach is reasonably required to satisfy the agency’s needs.  See Northrop Grumman Tech. Servs. Inc., B-406523, June 22, 2012, 2012 CPD ¶ 197 at 7.  A protester’s disagreement with the agency’s judgment concerning the agency’s needs and how to accommodate them, without more, does not show that the agency’s judgment is unreasonable.  General Electrodynamics Corp., supra

The record reflects that the agency decided “to procure the JCF requirement as one single integrator contract [in order] to minimize risk to the Government, scale adoption of AI across the [DOD], and to provide efficiencies in cyber security.”  AR, Tab 24, Decl. of JCF Branch Chief at 1.  The JCF branch chief explained that the integrator was a central function of the procurement that could not be separated from the other functions: 

[T]he Government possesses limited AI/ML expertise to properly assess and integrate those commercial or open source AI platforms for interoperability among the platforms themselves and government systems. This interoperability is critical to successfully scale development, testing, and delivery of AI applications.  Thus, by procuring the domain-specific AI platforms separately, the Government would assume an unacceptable level of risk by directly procuring AI platform services that may not:  (1) fully function with existing government and commercial systems; and (2) apply all AI use cases (e.g. natural language processing) across the breadth of [DOD] Mission domains.

Id. at 2 (internal citations omitted).  DISA also identified considerable efficiencies with a consolidated procurement in terms of obtaining ATOs for the AI/ML platforms, as follows:

Additionally, procuring each of the commercial AI platforms separately would create difficulty achieving Authorization to Operate (“ATO”) for each of the individual commercial AI platform services, delaying the ability to deploy those platforms across the [DOD].  This is because in order for software to be utilized by the [DOD], it must first be granted final ATO.  The final ATO has to be connected to a system.  If the Government bought each of the commercial AI platforms separately, they would each have to be given final ATO as separate, individual systems before being moved on to the integration platform.  However, if the software is brought on the single integration platform, the separate AI platforms would achieve final ATO from the operation of the integration platform.

Id.  DISA also found a benefit in the integrator providing a single point of customer support for all AI software, as opposed to separate customer support for each software.  Id. at 3.

At the hearing held in this protest, the JCF branch chief explained the importance of the integrator function:  “The JCF really has three fundamental categories of platforms.  We have the software integration platforms, we have the data platforms, and then we have the AI and ML platforms.”  Tr. at 99.  The integrator will “find the commercial platforms [with] those AI/ML capabilities . . . and evaluate those platforms [for] best practices.”  Id. at 99-100.  The agency’s need for an integrator rose from the complexity that “comes from the interdisciplinary nature of the JCF, where we cut across cloud engineering, data engineering, the [DOD] risk management framework.  And then sprinkling in the cognitive and AI expertise to be able to develop . . . the infrastructure to support” AI/ML technologies.  Id. at 74.  In addition, “the complicated nature . . . stems from the breadth and depth of making sure that we have the right processes in place to satisfy our authorizing officials to allow us to maintain continuous ATO.”[15]  Id

The protester has not shown that the agency’s decision to consolidate commercial and non-commercial requirements is clearly unreasonable.  While some of the individual components here would likely be considered commercial items if acquired individually, we think that the language of 10 U.S.C. § 2377 instructing the agency to consider modification of the agency’s requirements or modifications of the commercial items themselves does not, in every case, force an agency to have to separate out all potentially commercial items from a procurement.  See Analytical Graphics, Inc. v. United States, 135 Fed. Cl. 378, 430 (2017).[16] 

As stated above, the contracting agency has the primary responsibility for determining its needs and the best method of accommodating them.  Voith Hydro, Inc., supra.  On the record before us, the protester has not shown that the agency’s determination that its needs could not be met with a commercial item was unreasonable.  As such, this protest ground is denied. 

Procurement of Commercial Items in the JCF

The protester also argues that the RFTOP improperly excludes commercial solutions by using cost-reimbursement CLINs.  Protest at 26.  DISA contends that, to the contrary, the solicitation affirmatively provides for the integrator to seek and integrate commercial solutions.  COS/MOL at 33.

The FASA and the FAR require agencies to require prime contractors and subcontractors at all tiers to incorporate, to the maximum extent practicable, commercial items or nondevelopmental items as components of items supplied to the agency.  10 U.S.C. § 2377(b)(2); FAR 12.101.

Here, the RFTOP expressly provides that “[t]he contractor shall develop methodologies and procedures for acquiring, testing, and assessing commercial AI/ML tools and services,” i.e., for acquiring tools such as those developed by C3.ai.  RFTOP §  Similarly, the RFTOP states that “[t]he contractor shall develop and implement CI/CD [continuous integration/continuous delivery] pipelines that allow[] commercial and open source AI/ML tools to be rapidly vetted, through software assurance/supply chain provenance, and provisioned for use within the JCF DevSecOps environments.”  Id. §; see also id. § (“The contractor shall acquire, ingest, and condition commercial/public AI/ML data sets and make them available in the JCF.”). 

Furthermore, the RFTOP pricing worksheet contains a cost-reimbursement entry for software and licensing expenses.  AR, Tab 14j, RFTOP Pricing Worksheet.  The agency anticipated that the contractor could nevertheless incorporate commercial solutions.  In this regard, when asked by a potential offeror whether “a bid segregating commercial services from the CPFF [cost-plus-fixed-fee] CLIN is allowable,” DISA responded that “[t]he Government permits commercial subcontractors to be priced.  The Government’s arrangement with the Prime is under a Cost-plus-fixed-fee (CPFF) basis [and] the Government does not dictate contract types between the Prime and Subcontractor.”  AR, Tab 22k, RFTOP amend. 007, attach. 11, Question & Answer No. 104. 

In deciding that the procurement was non-commercial, the contracting officer acknowledged that commercially-available AI/ML tools were available and that some of the central functions of the integrator would be to assist vendors in ensuring that they have an ATO and that users across DOD can access the tools.  AR, Tab 11, Memorandum for Record at 1.  Finally, the JCF branch chief anticipates that the use of a central, private integrator will encourage the procurement of commercial products through subcontracting.  Tr. at 127:1-3 (“JCF emphasizes we want to take the barriers away from commercial items to be able to get into the [DOD] ecosystem.”).

On this record, we conclude that the solicitation does not exclude commercial solutions.  C3.ai, which offers a commercial platform with AI/ML capabilities, is not excluded in any way from participating in the solicitation as a subcontractor.  As noted above, a protester must demonstrate that it has been prejudiced by the agency’s actions.  Here, C3.ai does not show that it has been competitively harmed. 


Finally, we turn to the protester’s arguments regarding the scope of the SETI IDIQ contract.  C3.ai which does not hold a SETI contract, contends that the construction of a “platform-of-platforms” is outside of the scope of the SETI contract.  Protester Post-Hearing Brief at 4.  Specifically, the protester asserts that in task 2, “the [RFTOP] requires platform building and software development (which is not within the scope of the SETI Contract)[.]”  Protester Post-Hearing Brief at 4.  The protester argues that the JCF tasks 3, 4, and 5 are similarly outside of the scope of the PWS. 

The agency asserts that the SETI contracts were written broadly and reasonably include the services requested here.  In this regard, DISA contends that there are no material differences between the tasks described in the RFTOP and the work permitted under the SETI contract.  Agency Post-Hearing Comments at 15.  The agency also argues that the SETI contract was written broadly, with the intent of capturing future technologies.  Next, the agency maintains that, while many specific terms used in the JCF may not be in the SETI contract, they are well within the types of products and services described in the SETI PWS.  Id. at 14.  DISA asserts that, where a task order PWS uses terms that are not exactly the same as those in the underlying contract, the difference does not automatically render those tasks outside of the scope of the IDIQ contract.  Id. at 13 (citing Morris Corp., B‑400336, Oct. 15, 2008, 2008 CPD ¶ 204).

CICA requires agencies to obtain “full and open competition” in procurements through the use of competitive procedures.  10 U.S.C. § 2304(a); see also 41 U.S.C. § 3301(a).  Task orders that are outside the scope of the underlying multiple-award contract are subject to the statutory requirement for full and open competition set forth in CICA, absent a valid determination that the work is appropriate for procurement on a sole-source basis or with limited competition. 10 U.S.C. § 2305(a)(1)(A)(i); 41 U.S.C. § 3301; see American Sys. Grp., B-418469, Apr. 7, 2020, 2020 CPD ¶ 140 at 2. 

In determining whether a task or delivery order is outside the scope of the underlying contract, and thus falls within CICA’s competition requirement, our Office examines whether the order is materially different from the original contract, as reasonably interpreted.  Nuance Commc’ns, Inc., B-418106, Jan. 8, 2020, 2020 CPD ¶ 23 at 4-5.  Evidence of a material difference is found by reviewing the circumstances attending the original procurement; any changes in the type of work, performance period, and costs between the contract as awarded and the order as issued; and whether the original solicitation effectively advised offerors of the potential for the type of orders issued.  Id.  In other words, the inquiry is whether the task order is one which potential offerors should have reasonably anticipated.  Id.; Symetrics Indus., Inc., B-289606, Apr. 8, 2002, 2002 CPD ¶ 65 at 5.  

Where there is a logical connection between a broad scope of work in an IDIQ contract and the services procured under a subsequent task order, potential offerors are on notice that such logically-connected services are within the scope of the IDIQ contract.  Id.; see also Morris Corp., B-400336, Oct. 15, 2008, 2008 CPD ¶ 204 at 5-6.  IDIQ statements of work need not include specific references to every type of task, particularly where the challenged task reflects an insignificant fraction of the total value of the IDIQ contract.[17]  Specialty Marine, Inc., B‑293871, B‑293871.2, June 17, 2004, 2004 CPD ¶ 130 at 6; Outdoor Venture Corp., B-401628, Oct. 2, 2009, 2009 CPD ¶ 200 at 3.

Based on our review of the record, we find that the SETI contract encompasses the services requested by the RFTOP.  The RFTOP summarizes its scope as follows:

(1) Engineer and Architect the JCF.  Encompasses engineering and architecture of all aspects of the JCF including overall systems engineering and architecture, cloud engineering and architecture, engineering and architecture of AI/ML tools and services within the JCF, data and data services engineering and architecture, DevSecOps engineering and architecture, and security engineering and architecture.

(2) Build, Secure, and Enhance the JCF.  Encompasses all aspects of building, securing, and enhancing the JCF as a software platform using modern, agile, CI/CD practices. [ . . .]

(3) Operate and Defend the JCF.  Encompasses all aspects of operating, maintaining, and defending the JCF, including managing cloud capacity, provisioning instances of the JCF, operating and maintaining JCF instances, monitoring and defending the JCF, onboarding new users and projects into the JCF, and logging and auditing JCF activities.

(4) Develop and Implement Standardized [DOD]AI workflows, Patterns, and Best Practices.  Encompasses developing and maintaining a standardized set of [DOD]-specific AI workflows, development patterns, and best practices that are designed to be appropriate for the bulk of the [DOD] AI community.  The workflows, patterns, and best practices will help break down barriers to AI development in the [DOD].

(5) Provide Project and Product Management. This task encompasses all aspects of managing the JCF Prime Integrator program and all aspects of providing product management oversight of the JCF.

RFTOP at 3-4.  The PWS provides additional details on all of these tasks.  Id. at 4-40. 

The SETI contract states that it was written broadly in order to encompass future endeavors and technologies whose specifics were unknown at the time for award.  SETI Contract at 20.  For example, the SETI contract’s identified purpose was to “provide a suite of information technology (IT) engineering services, expertise, and support in the planning, research, development, integration, and implementation activities for future, proposed, current, and legacy [] ([DOD]) and [] (DISA) IT capabilities, services, and systems.  It will deliver complex IT engineering support services throughout the entire acquisition lifecycles of [DOD] and DISA solutions, systems, and system components.”  Id.

With regard to “Cloud Engineering,” the SETI contract states that “the Contractor shall implement the cloud model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) . . . .”  Id. § 6.8.5.  Furthermore, “[i]n support of this area, the [SETI] Contractor shall perform activities including, but not limited to:

• Develop and augment cloud computing technologies to optimize cost savings combined with increased IT agility

  *  *  *  *  *

• Engineer cloud solutions

• Develop and engineer secure and reliable IT computing services within the cloud environment.

• Integrate new applications into cloud environments, and transition existing applications into cloud environments

• Engineer a solution to integrate various platforms and eliminate the need for multiple interfaces to manage cloud environments

• Integrate other components from other systems and/or cloud environments, including commercial products and systems

  *  *  *  *  *

• Apply security requirements across crosscutting layers of the solutions architecture of all layers in the cloud environment (i.e., end-to-end security that ranges from physical security to application security).

• Conduct risk assessment and viability studies for transitioning various applications into the cloud environment

Id. § 6.8.5 Subtask 5 – Cloud Engineering. 

The breadth of the SETI contract’s scope is apparent both in the enumerated activities as well as the provision that allows for activities beyond those specified.  Id.  To the extent that the protester argues that the SETI contract provides only for support for a single platform or system, and not a platform-of-platforms, we disagree.  The contract as to its scope states that “SETI offers an array of research and development, and critical technical disciplines core to engineering, delivering, and maintaining [DOD] and DISA IT products and capabilities.”[18]  Id. § 5 – Scope.  In addition, SETI “encompasses legacy, current, and future [DOD] and DISA capabilities, services, systems, and systems-of-systems (SoS) throughout the acquisition and system engineering lifecycle.”  Id.  The SETI contract is instructive as to its own interpretation, in that “[w]here applicable, the term ‘system’ can refer to a single system, collection of systems, subsystems, and SoS, together within its environment.”[19]  Id. 

As to services, the SETI contract also requires the contractor to “provide support and services for the full range of technology areas that are based on Artificial Intelligence and Signal Processing . . . .”  Id. § 6.8.8 Subtask 8 – Cognitive Computing.  Second, the SETI contractor “shall provide systems engineering support and services to the application of new technologies, which may not yet be mature at the time of Contract award, to improve and expand DISA services to meet future mission needs . . . .”  Id. § 6.8.9 Subtask 9 – Emerging Technologies.

We find that these descriptions in the SETI contract reasonably encompass the tasks described in the RFTOP.  The scope of the SETI contract is broad.  Although the JCF may not use the same words and phrases as the SETI contract, that does not itself render the task order functions materially different from the contract scope.  Nuance Commc’ns, Inc., supra, at 4-5 (broad PWS in underlying IDIQ reasonably encompasses work describe in task order, even where task order uses specific terms).  We conclude there is a logical connection between the contract and the task order, and therefore find no basis to sustain the protest here. 

The protest is denied.

Thomas H. Armstrong
General Counsel


[1] The program is called the Joint Common Foundation (JCF).  Agency Report (AR), Tab 12, Acquisition Plan at 1. 

[2]  Part 12 of the FAR establishes procedures for agencies to use to determine whether commercial items or nondevelopmental items are available that could meet the agency’s requirements.  See, e.g., FAR 12.101.  

[3] C3.ai is not a SETI contract holder.  In 2019, the agency also awarded 25 SETI contracts to small businesses.  Contracting Officer’s Statement and Memorandum of Law (COS/MOL) at 10.  Only the suite of SETI contracts awarded to large businesses is relevant to this protest.  Although the SETI contracts were awarded to each large business as a separate contract, for simplicity this decision refers to them as a single SETI contract.  Citations to the SETI contract are to the document at tab 2a of the agency report.

[4] An authorization to operate--sometimes called authority to operate--is the official management decision given by a senior organizational official to authorize operation of an information system and to explicitly accept the risk to organizational operations (including mission, functions, image, or reputation), organizational assets, individuals, other organizations, and the nation based on the implementation of an agreed-upon set of security controls.  DATA Act: OMB and Treasury Have Issued Additional Guidance and Have Improved Pilot Design but Implementation Challenges Remain, GAO-17-156at 28 n.37 (2016) (citing the definition for ATO set forth in National Institute of Standards and Technology Special Publication 800-37).  The contractor will administer a continuous ATO by ensuring that the changes made to the software within the JCF platform do not invalidate the security of the software and the software is still permitted to operate within the DOD networks.  Hearing Transcript (Tr.) at 23-25.

[5] These six large businesses were [DELETED].  AR, Tab 10, April Market Research Report at 5-6, 8-9.  Five of the six entities intend to use partners or subcontractors to complete at least some of the work.  AR, Tab 6g, [DELETED] Sources Sought Response at ii; Tab 6r, [DELETED] Sources Sought Response at 2; Tab 6t, [DELETED] Sources Sought Response at 4, 10; Tab 6w, [DELETED] Sources Sought Response at 2; and Tab 6zb, [DELETED] Sources Sought Response at 12.

[6] C3.ai appears to offer only a platform, not systems integration services.  See AR, Tab 6j, C3.ai Sources Sought Response at 9 (“C3.ai also maintains partnerships with defense system integrators such as Raytheon, hardware providers such as Intel, edge analytics companies such as FogHorn, and Systems Integrators such as IBM and West Monroe Partners.”); at 2 (“In partnership, we can provide a comprehensive software and support solution to serve as the foundation for the JAIC JCF.”) (emphasis added).  In addition, at the time it submitted its response to the agency’s sources sought notice, C3.ai did not have an ATO, having first applied for one in November 2019.  Id. at 5. 

[7] According to DISA, the Defense Innovation Unit “accelerates the adoption of commercial technology into the U.S. military to strengthen national security.”  AR, Tab 1, Summary of 2018 DOD AI Acquisition Strategy at 13.

[8] The document is dated March 12 but the contracting officer’s electronic signature is dated April 29.  The agency states that the delay between drafting and signing was “due to administrative discussions regarding the appropriate format and signatory to the [document].”  Agency Supp. Resp. at 23 (internal quotations omitted).  DISA asserts that the document’s contents describe subsequent events that establish that it “was created prior to the April 2, 2020 email in a different format.”  Id. at 7.  The agency’s explanation is reasonable and consistent with the record, and we have no reason to believe that the contracting officer reached the conclusions therein after the RFTOP was issued.

[9] It appears that the new contracting officer did not know that the temporary contracting officer had previously signed a version of the market research report in March.  AR, Tab 27, Email from DISA Contracting Officer to DISA JCF Branch Chief, Apr. 2, 2020, at 1.  

[10] Citations to the RFTOP are to the RFTOP performance work statement (PWS) at tab 22a of the AR. 

[11] The revised market research report was signed on April 2, but the contracting officer stated that she and another DISA employee noticed the “discrepancy” in the conclusion regarding the non-commercial nature of the procurement on March 23 and revised the document at that time.  AR, Tab 10, April Market Research Report at 10; Tab 27, Email from DISA Contracting Officer to JCF Branch Chief, Apr. 2, 2020, at 1. 

[12] Because the value of the task order here is over $25 million, this procurement is within our jurisdiction to hear protests related to the issuance of orders under multiple-award IDIQ contracts issued by military agencies. 10 U.S.C. § 2304c(e)(1)(B).

[13] The protester also raises collateral protest grounds.  We have considered all of the protest grounds and, although not all are discussed here, none provides a basis to sustain the protest.  For example, the protester asserts that the RFTOP is outside the scope of the SETI contract because it will likely exceed the value of a single SETI contract, which C3.ai argues is capped at $250 million based on information in the government’s System for Award Management.  Protest at 13 (citing exh. 8, Summary of SETI Contract), 19; but see Comments at 56 (“C3.ai is not challenging the reasonableness of the IGCE as a separate protest ground[.]”).  However, even if the protester’s estimate (disputed by the agency) is accurate, the SETI contract permits the agency to reallocate unspent amounts between SETI contract holders “if necessary,” with no limitation on what qualifies as “necessary.”  SETI Contract at 2.  As adjustment of contract value is permitted and C3.ai does not claim that the JCF task order will exceed $7.5 billion (or even $300 million, i.e., the sum of the not-to-exceed amounts for the cost-reimbursement CLIN for the four performance periods), we find no merit in this protest ground. 

[14]The protester challenges the agency’s decision to “bundle” its requirements.  See, e.g., Comments at 65.  The FAR defines “bundling” as a consideration related to small businesses.  FAR 2.101.  The protester neither claims to be a small business nor that the services in the RFTOP were previously provided by small businesses.  To the extent that the protester means “consolidation” or “consolidated requirement,” the FAR defines these terms as a single procurement for two or more requirements previously provided for the agency in question under separate contracts, “each of which was lower in cost than the total cost of the contract for which offers are solicited.”  Id.  The protester does not allege that the various services specified here were previously provided to DISA

[15] The protester acknowledges the complexity of the requirement.  C3.ai submitted with its comments a declaration from its outside consultant, who was admitted to the protective order.  See Electronic Protest Docket System Entry No. 25.  The consultant explained that, in his professional opinion, “[t]he [RFTOP] calls for work that requires highly specialized subject matter expertise for the creation of cutting-edge software-reliant capabilities based on hundreds of AI Workspaces, which allow one or more AI developers to develop, test, secure, and deliver AI capabilities and services.”  Comments, exh. 1, Consultant Decl. ¶ 11. 

[16] The agency asserts that our Office should defer to its determination about the commercial nature of the procurement.  COS/MOL at 22 (citing Palantir USG, Inc., B‑412746, May 18, 2016, 2016 CPD ¶ 138 at 4).  The protester maintains that DISA failed to properly consider commercial solutions and that the resolution of the protest should be guided by the opinion issued by the Court of Appeals for the Federal Circuit.  Protest at 25 (citing Palantir USG, Inc. v. United States, 904 F.3d 980 (Fed. Cir. 2018).  While we acknowledge the decision of the Federal Circuit in Palantir, we think that the more relevant fact pattern is found in Analytical Graphics, Inc., 135 Fed. Cl., supra, where the agency required a solution with commercial and non-commercial aspects.  Id. at 393.  The protester in Analytical Graphics similarly argued that the agency was obligated to bifurcate its solicitation into separate procurements.  Noting the “extremely fact-based commercial availability issues,” the court concluded that the agency had articulated a reasonable basis for combining its commercial and non-commercial needs.  Id. at 397, 434. 

[17] As noted above, the protester does not challenge the agency’s formulation of the IGCE, and similarly does not allege that any of the allegedly out-of-scope tasks are significant in comparison to the overall contract value.  Cf. Leupold Stevens, Inc., B‑417796, Oct. 30, 2019, 2019 CPD ¶ 397 at 7.

[18] The RFTOP states that while “[t]he long-term objective is to build and operate the JCF in the [DOD]’s future [DOD] Enterprise Cloud capability provided by Joint Enterprise Defense Infrastructure (JEDI). . . .  at this time, the JEDI acquisition is not finalized and JEDI may not be available for the JCF until well into the first or second year of the contract.  Therefore, in the interim, the Government will use this contract as a mechanism to procure cloud capacity and services from [DOD] Enterprise Cloud Service Providers.”  RFTOP at 3. 

[19] This broad definition of the term “system” provided in the SETI contract is consistent with the agency’s understanding and use of the term generally, as the record reflects that the agency uses a number of different terms interchangeably.  The JCF branch chief testified that “systems” and “software systems” are “interchangeable terms” and are also “equivalent” terms in the context of the JCF.  Tr. at 49.  Furthermore, the “JCF is software,” and “[a] platform is software.”  Id. at 49-50.  Finally, “[i]n the context of JCF, all platforms are software,” and, in the context of the JCF, all systems are platforms.  Id. at 50-51.