NeedipeDIA CSO HHM402-23-SC-0002 CALL - TFS Amendment 2.pdf

UNCLASSIFIED

UNCLASSIFIED

COMMERICAL SOLUTIONS OPENING (CSO) NEEDIPEDIA CALL

HHM402-23-SC-0002

Call Announcement CSO-0025-0001

CSO CALL TITLE: Fiscal Year 2025 Task Force SABRE (TFS) Artificial Intelligence and Machine Learning (AI/ML)

CONTRACTING POINT OF CONTACT:

Mr. Il Jun Yun, Unclassified Email: iljun.yun@dodiis.mil;

REQUIREMENT DESCRIPTION:

1. Task Force SABRE (TFS) is a limited duration organization charted to provide the Defense Intelligence Agency (DIA) with foundational, enterprise-wide Artificial Intelligence (AI) capabilities while accelerating adoption. TFS will establish conditions which ensure widespread adoption of AI in DIA, ensuring the Agency outpaces our adversaries and provides warfighters with decision advantage in the era of Artificial General Intelligence and Agentic AI. The TFS seeks revolutionary or innovative AI/ML prototype project, as defined in 10 U.S.C 4022(e)(5), focusing on the next generation of capabilities and platforms to outpace strategic competitors and advance enterprise integration. Competitive submissions will develop and demonstrate new AI/ML capabilities currently not resident in legacy systems, but may significantly improve existing capabilities, and/or will apply techniques developed for commercial or non-intelligence applications to current intelligence gaps in an ingenious manner (i.e., innovation). Potential projects will be considered if the benefits clearly outweigh the cost and time involved to mature innovative prototype into functioning AI/ML capabilities. Successful projects are those that meet successful completion criteria established in each award.

2. Process Overview: This CALL begins the process of identifying and prioritizing potential prototype projects for execution in Government fiscal year (FY) 2025. The TFS AI/ML project selection process involves two (2) rounds of evaluation of potential projects that advance the AI/ML technical gaps (TG) detailed below. During the first round of evaluation, the Government requests white papers proposing innovative prototype projects, which address technical gaps as derived Task Force SABRE’s DIA-wide requirements analysis, with a goal of receiving new concepts and participation from a variety of new sources of innovation. The white paper must also include a rough cost estimate with projected labor categories, labor cost, a schedule for delivery with milestones, projected material purchase costs, and breakout of specific other direct costs (ODC). During the second round of evaluation, for white papers selected from the first round of evaluation, the Government requests technical and cost/price proposal that describes technical approach and cost/price elements in details.

a. In reference to the DIA NeedipeDIA CSO HHM402-23-SC-0002. Page 12, Part II, Section E.1:

UNCLASSIFIED

This CSO requires interested parties to separately review https://www.dia.mil/Business/NeedipeDIA/ prior to completion and submission of any White Papers. Interested parties…self-identify…the single most appropriate NeedipeDIA listing that is most closely associated with their white paper concept/idea.

To the extent possible, interested parties should also self-identify the most appropriate Technical Gap listed in Section 3.

b. In reference to the DIA NeedipeDIA CSO HHM402-23-SC-0002, SUBMISSION OF WHITE PAPERS: Parties interested in engaging DIA with innovative topics of interest shall submit a White Paper to DIA for review, prior to submission of any proposal related information. All unclassified white papers shall be submitted via email to Ideas2Action@dodiis.mil. Contact Ideas2Action@dodiis.mil for classified white papers submission instructions.

c. Questions/Inquiries – Interested parties may submit questions or inquires through the listing for this CALL on SAM.gov or via email. Answers to questions/inquiries will be provided through SAM.gov. Please be specific in your requests for clarity to the CSO.

The Questions/Inquiries period begins from the release of this CALL through 15 April 2025. Please submit questions/inquiries on the listing for this CALL on SAM.GOV or to the following email addresses iljun.yun@dodiis.mil (unclassified).

d. For this CALL, additional instructions and guidance for the white paper are included below.

This ad-hoc CALL includes descriptions of technical gaps to provide additional need area clarification. Offerors can tailor white papers to provide innovative solutions for technical gap(s) identified in this CALL. The selection criteria are based on the technical approach to fill critical gaps, the timeline and costs associated with the proposed solution. White papers that do not meet the below requirements will not be considered. The prototype project described in the white paper must clearly articulate relevancy to at least one AI/ML technical gap described in this call. The white paper shall provide a rough estimate of cost, including expected cost and duration of the project. Labor, Materials, ODC and timeline must be clearly identified. White papers submitted via any other mechanism will not be accepted or reviewed under this CSO CALL. No information submitted by an interested party in response to this CSO will be returned. In addition, the cost of preparing White Papers in response to this CALL is not considered an allowable direct charge to any resulting procurement instrument.

i. White Paper Format (2 Page Limit): The format for submission of White Papers is intended to be flexible, but all submissions should adhere to the following general guidelines:

UNCLASSIFIED

1. Maximum of two (2) pages in length, not including the coversheet.

2. Coversheet as downloaded in Attachments and completed in its entirety

3. Font Type should not be smaller than 12 pt, although smaller font type may be used for figures, tables, and charts

4. All White Papers must be written in English

5. All White Papers should contain the following key elements

a. Part 1 Coversheet: (does not count against 2-page limit)

b. Part 2 Executive Summary: The Executive Summary should address the following questions: Clearly identify your technical approach. What are you trying to do? How is it currently done? What innovative claims is the interested party making? Why is it beneficial to pursue proposed solution in comparison to commercially existing capability? Who cares?

What are the risks? What’s your plan? What are the deliverables and successful completion criteria? What’s your progress? Identify the primary Government stakeholder. What facility(ies) will this work take place in?

c. Part 3 Other Applicable Information: Provide any additional information pertinent to your White Paper submission.

Examples include, but are not limited to, the following: your desired classified or proprietary discussions with the government, foreign participants or organizations that are involved in the effort, if you anticipate subcontracting or partnering with others to accomplish the initiative in question, plan to address the need for Government equipment or Government data, rough estimate of cost (Labor, materials, and ODC’s included).

d. Part 4 Technical Gap: Clearly identify at least one AI/ML Technical Gap listed in Section 3.

ii. Offerors are strongly encouraged to identify a potential sustainment plan beyond the initial pilot within the white paper. A sustainment plan should include approximate costs for personnel, technology, and maintenance.

iii. Offerors are encouraged to identify methods to integrate their solutions to DIA networks, infrastructure, and common IT services.

iv. Offerors are encouraged to identify methods to provide cyber security related documentation required for testing and piloting their solution on DIA networks.

e. A multi-discipline working group of technical experts will evaluate the white papers.

White paper submissions demonstrating innovative potential to address high-priority Technical Gaps will be selected.

UNCLASSIFIED

3. Technical Gaps: Before submission of any white papers, potential offerors are to review the topics listed below and self-identify (in both the coversheet and in the white paper), the single most appropriate Technical Gap that is most closely associated with their white paper concept or idea, or where the preponderance of benefit is most likely to occur, even if their particular concept or idea potentially covers more than one topic area. Potential offerors may submit multiple white papers covering various concepts; however, offerors are encouraged to submit quality over quantity as each submission will be evaluated individually. There is no selection quota by topic, nor maximum limit for submissions to be considered for each topic. Note: CSO HHM 402-23-SC-0002 uses the term "Need" interchangeably with "Topic;" “Research Areas of Interest”.

a. TG-1: Enterprise Multi-Model Large Language Model Application. Large Language Models (LLM) are a type of artificial intelligence designed to process large amounts of data and generate human-like responses to natural language inputs. The enterprise application will take advantage of already available LLMs but must include the capability to integrate new models as they become available and must be fully contained on the DIA JWICS environment. Solutions must include an intuitive User Interface which includes the ability to select different LLMs, an output cache for prompt optimization, the ability for users to upload documents into temporary memory for incorporation into the output cache, and an output display. Solutions must be scalable to a large distributed user base, of greater than 2000 concurrent users, and incorporate necessary user authentication and cyber information security features as required by DIA CIO for JWICS deployment, and NIST 800-53, and ICD 503 controls. Solutions must include incoming and outgoing API capabilities. Solutions must be easily extendable to incorporate current and future advanced features such as Retrieval-Augmented Generation (RAG), media generation, and AI Agents.

b. TG-2: Multi-Source Data Retrieval and Pre-Processing. The Intelligence Community (IC) possess large, geographically separated data holdings. Solutions must retrieve these data from disparate sources, pre-process based on the needs of the relevant analysis or AI tool, and effectively manage data storage. Solutions must efficiently manage high volume, high velocity data. Solutions must have incoming and outgoing API capabilities for efficient integration into existing and emerging workflows. Solutions must include all cyber and information security features required by DIA CIO for JWICS deployment.

Solutions must account for data stores at multiple classification levels. Solutions must support fine-grained data entitlements management. Solution must either include or have the ability to incorporate a cross-domain solution for moving data between classification levels. Solutions must include the ability to construct and store vector databases in support of Retrieval-Augmented Generation implementation into an LLM application.

c. TG-3: AI-Enabled Document Generation and Assessment. Generating standardized reports and documentation creates a lot of administrative overhead on organizations.

LLMs possess the capability to process documents and return the data required to populate the reports but not in a specific format. Solutions shall ingest and parse documents using domain-specific knowledge, configure and send data via API to an

UNCLASSIFIED

LLM application, receive and parse the LLM response using domain-specific knowledge, and return the data in a standardized form. Solutions shall have the ability to assess the LLM response based on domain-specific knowledge and provide a feedback mechanism to the LLM based on said assessment. Solutions shall have both automated and manual data retrieval and storage capabilities. Solutions shall include all cyber and information security features required by DIA CIO for JWICS deployment.

4. Critical Requirements: Offerors responding to this Call must satisfy all of the following top-level requirements:

a. Clearance and Classification. Offerors must be able to work with and store classified information at the level appropriate for the work to be performed, possibly up to and including TOP SECRET // SCI. Therefore, the potential offeror must have an identified workforce with the appropriate security clearances (or the ability to obtain the appropriate clearances in time for execution of the project), and must at the time of proposal submission have access to a facility to store and process material at the level appropriate for the work to be performed, possibly up to and including TOP SECRET // SCI.

b. SAM.gov. Offerors must be registered in the System for Award Management (SAM) at https:///www.sam.gov. Offerors shall have no active exclusions nor have criminal, civil or administrative proceedings listed within their SAM.gov record.

PARTICIPATION: The TFS AI/ML CSO CALL is open to ALL interested parties and partnerships, including Federally Funded Research and Development Centers (FFRDCs) (e.g., National Laboratories, Centers), government/military offices, Service Laboratories (e.g., Air Force, Navy, and Army Research Labs), academia (e.g., colleges, universities, centers), and industry (e.g., traditional/nontraditional defense contractors, small business). (As noted in HHM402-23-SC-0002, "Non-U.S. organizations and/or individuals may participate to the extent that such participants comply with any necessary nondisclosure agreements, security regulations, export control laws, and other governing statutes applicable under the circumstances.")

ANTICIPATED FUNDING: Funding for TFS varies by fiscal year (FY) and the number of proposals funded. Funding amounts awarded will depend on the availability of program funds and on the number and applicability of submissions received.

PERIOD OF PERFORMANCE: Proposals that will enhance capabilities within six (6) months are preferred, although longer concepts that use option period(s) will be considered.

ANTICIPATED TYPE OF AWARDS: The Government Contracting Officer / Agreements Officer, at his/her sole discretion, reserves the right to negotiate other appropriate award instruments which may include acquisition contracts, other transactions (OT), or interagency agreements/transfers. Offerors will note their requested funding instrument on their white paper coversheet and in their proposal.

UNCLASSIFIED

ANTICIPATED NUMBER OF AWARDS: The number of projects funded will depend on the availability of program funds and on the number and applicability of submissions received.

ANTICIPATED AWARD DATE: Projects selected for the TFS AI/ML CSO CALL will begin receiving funding and start execution with a target of June 2025 for the first projects based on prioritized selection, with other projects to follow as soon as the funding process allows. This timing depends on government funding preparation timelines. Rather than issuing another CALL, TFS may retain non-selected proposals from this CALL for potential funding if subsequent funds become available. If TFS selects this option, proposers will be notified and given an opportunity to update technical, management, and cost proposals.

TIMELINE OF EVENTS: Questions concerning this call letter will be accepted by the Government’s Contract Specialist Il Jun Yun (iljun.yun@dodiis.mil) with an end date of 15 April 2025 at 17:00 ET.

TIMELINE AND OFFEROR DEADLINES:

Target TFS AI/ML White Paper Submission and Evaluation Milestone:

CSO CALL posted in SAM.gov White Paper submission and Questions/Inquiries period begins Final date to submit Questions/Inquiries Response to Questions/Inquiries will be posted in SAM.gov Window closes at 11:59pm ET for White Paper Submissions

07 April 2025 07 April 2025 15 April 2025 18 April 2025 07 May 2025 12 May 2025 Offerors receive Government intent to proceed or not-to-proceed to follow-on engagements Window for follow on engagement from Government opens Window for follow on Engagement from Government closes

14 May 2025 16 May 2025 19 May 2025 Offerors receive Government intent to proceed or not-to-proceed to

Request for Technical and Cost/Price Proposals Request for Technical and Cost/Price Proposals released Proposals due by 11:59pm ET

21 May 2025 30 May 2025 23 June 2025 Award date (subject to change based on availability of funds)

WHITE PAPER DUE DATE AND TIME: The due date for white papers submitted in response to this CALL is no later than 07 May 2025 at 2359/11:59 pm ET. White papers received after the due date and time will not be considered for selection.

WHITE PAPER EVALUATION CRITERIA: In addition to the criteria described in the baseline CSO, technical papers will also be evaluated based on responsiveness to the following technical metrics:

A. Degree in meeting the AI/ML need requirements.

B. Speed of implementation.

C. Innovative aspect of proposed solution and benefit of pursuing proposed solution in comparison to commercially existing capability.

d114987 Highlight d114987 Highlight d114987 Highlight d114987 Highlight d114987 Highlight d114987 Highlight d114987 Highlight d114987 Highlight d114987 Highlight

UNCLASSIFIED

D. Proposed project development methodology.

E. Experience with proposed technological area.

F. Feasibility of implementing solution within DIA networks G. Ability of the solution to satisfy DIA cyber security requirements to obtain an Interim

Authority to Test, up to an Approval to Operate.

H. If applicable, description of why proposed sub-contractors were selected, significance of sub-contractor’s participation and what tasks they are to perform.

DELIVERABLE ITEMS: Generally, periodic deliverable requirements include written technical and financial progress reports (monthly to quarterly), and in-person/virtual Technical & Financial Progress Review briefings (quarterly to biannually), as agreed between TFS and the performer for each project. Final deliverables will include a written scientific/technical report and briefing describing problem, intended outcome, relevance, approach, findings/development, etc. In case of software development, user feedback and adoption of improvement to enhance user interface is expected. Final deliverable end-items will vary somewhat based on the nature and scope of the prototype project proposed and formal reporting or deliverable requirements will be specified in each projects’ Statement of Work (SOW). Examples include but are not limited to: physical prototypes and hardware/software items developed from prototype project;

test and evaluation report of final prototype product; development of algorithms and software solutions may reach sufficient maturity for deployment and integration into operations during the funding period.

SUCCESSFUL COMPETION CRITERIA: In case of prototype OTA award, successful completion criteria will be established at the time of award based on proposed solution.

CALL AMENDMENTS: Offerors should monitor http://www.sam.gov for any additional notices to this CALL that may permit extensions to the proposal submission date or otherwise modify this announcement.

OTHER RELEVANT INFORMATION: None

GOVERNMENT FURNISHED EQUIPMENT: Government furnished property, equipment, or facilities needed for this requirement should be specified in the proposal.

DOD SECURITY AND DATA DISSEMINATION: Offerors must mark their proposals with the restrictive language stated in FAR 15.609(a), and in accordance with all applicable security guides. Submissions classified SECRET Collateral and below shall be addressed to the Contracting Point of Contact listed above. If deemed necessary by the Government, a form DD254, CONTRACT SECURITY CLASSIFICATION GUIDANCE will be issued with the subsequent contracting actions. It is a material condition of this arrangement that the candidate shall not use, or allow to be used, any information concerning this effort for advertisement, publicity, reference, or any other public purpose without specific approval of the Government sponsor. This includes statements in marketing tools, literature, or personal resumes. The contractor must also gain approval from the Government sponsor for any technical publications based on the research conducted under this contract.

UNCLASSIFIED

USE OF NON-GOVERNMENT ADVISORS: Offerors are advised that data included in white paper, and technical/cost proposals submitted to the Government in response to this CSO call may be released to support contractors or Federally Funded Research and Development Center (FFRDC) personnel for review and analysis. Any such personnel will strictly serve in an advisory capacity only to the Government evaluators and will provide the TFS with independent review and analysis on AI/ML technical matters and on accomplishment of objectives, tasks, and deliverables as proposed or required by TFS. All government and non-government participants in the review process are required to execute government nondisclosure agreements and personal/organizational statements of financial interest, as appropriate. By submission, offerors agree that their material may be reviewed by support contractors.