RFI - DAU Generative AI Based Acquisition Planning and Decision Capability.pdf

REQUEST FOR INFORMATION (RFI)

DAU Generative AI Acquisition Planning and Decision Capability

A. Purpose: This is a Request for Information (RFI). This is NOT a solicitation for proposals, proposal abstracts, or quotations. The Department of Defense (DoD), Washington Headquarters Services (WHS), Acquisition Directorate (AD) is seeking firms doing business in North American Industry Classification System (NAICS) codes 541511 (Custom Computer Programming Services) and 541715 (Artificial intelligence research and development laboratories or services). The purpose of this Request for Information (RFI) is to solicit white papers from industry detailing concepts for implementing a generative AI-based application that enhances the planning and decision-making processes for defense acquisition program teams and their stakeholders.

This request is issued for the purpose of conducting market research to 1) determine what generative AI concepts or solutions currently exist in the marketplace that can provide planning and decision-making support to the defense acquisition workforce; 2) determine if there are near-term generative AI solutions in development and the level of maturity; 3) determine what cross-functional or single functional solutions are being explored or have developed solutions using generative AI; and 4) receive information and feedback on vendor interest.

B. Introduction/Background: This is a Defense Acquisition University (DAU) Request for

Information (RFI). DAU provides the global learning environment for the acquisition workforce through certification courses, job-specific training, web events, credentials, workshops, consulting, and job support tools. These resources provide career-long access to learning assets tailorable to the specific needs of acquisition workforce members, program teams, and their stakeholders. To improve its capability to provide a world class, cutting edge global learning environment, DAU is conducting introductory market research into generative Artificial Intelligence (AI) / Machine Learning (ML) technologies that can provide a capability to enhance complex acquisition planning and decision making for acquisition professionals. Defense acquisition program teams engage in complex planning and decision-making processes that require the collaboration of functional area experts and a wide range of stakeholders. Legacy acquisition planning & execution processes rely on dated technology and manually intensive methods resulting in unexploited information and data resources, missed collaboration opportunities, wasted labor hours, and other inefficiencies. Generative AI technologies have the potential to streamline these processes, improve efficiency, and enhance data-centric planning and decision-making.

C. Objective: The complexities of the interacting decision support systems for requirements, resources, and acquisition (these interacting processes are often referred to as Big A acquisition) create built in risk and often lead to unsatisfactory outcomes regarding cost, schedule, and performance. Program teams struggle under the burden of an unrelenting flood

DEPARTMENT OF DEFENSE

WASHINGTON HEADQUARTERS SERVICES

1155 DEFENSE PENTAGON

WASHINGTON, DC 20301-1155

ACQUISITION DIRECTORATE

of reports, program artifacts, technical data, and layers of process compliance requirements.

Creative and critical thinking often compete against the “Tyranny of the Urgent” which commonly leads to less-than-optimal decision-making. Ultimately these factors compound and lead to delays in delivering the right capabilities to the warfighter and other end-users.

Accordingly, DoD acquisition program teams and their stakeholders require new capabilities to greatly accelerate acquisition planning and decision-making, enhance creative and critical thinking, streamline data and information integration, and improve the sense-making of that data and information. Generative AI has opened an opportunity not only to accelerate and improve generation of program documentation and plans but also the opportunity to greatly enhance the capability of acquisition program teams to conduct the following tasks:

Anticipate and prepare for potential challenges and disruptions. Generative AI can help ideate more diverse and realistic scenarios by analyzing vast amounts of data, including technological trends, program data, intelligence data, technical data, and industrial base information. It can help ideate plausible alternative future scenarios to test assumptions and identify potential challenges and disruptions that may not have been considered by program teams.

Identify and exploit opportunities in the market/industrial base. By processing and analyzing large datasets, generative AI may uncover hidden patterns, correlations, and opportunities in the defense and commercial markets and generate insights that may enhance strategies.

Develop contingency plans and risk mitigation strategies. Generative AI could assist in generating a wide range of potential responses and strategies to address various identified risks and challenges and provide data-driven suggestions for contingency plans and risk mitigation strategies, helping program teams be better prepared for different scenarios.

Foster creative problem-solving. Generative AI can stimulate creative thinking by generating novel ideas, solutions, and approaches to problems encountered during program planning and execution. By providing a diverse set of AI-generated suggestions, it can inspire participants to think outside the box and explore innovative solutions.

Enhance decision-making capabilities and strategic thinking. Generative AI can support decision-making by providing near real-time analysis and insights. It can process complex data, evaluate the potential outcomes of different decisions, and offer data-driven recommendations to guide strategic thinking.

Improve communication and collaboration among team members. Generative AI can facilitate communication and collaboration by providing a common language and framework for discussing complex scenarios and strategies. It can generate visualizations, summaries, and reports that help team members better understand and communicate the insights.

Increase efficiency and accuracy: Reduce the time required for drafting program documentation and increase the accuracy of incorporating authoritative data and applicable statutory and regulatory requirements.

Ensuring key data and or opportunities are not overlooked: Review and evaluation of the acquisition deliverables to ensure that key data requirements were addressed, and modern acquisition opportunity paradigms were considered.

The ideal concept will apply generative AI to enable cross-functional program teams to rapidly generate and wargame alternative acquisition scenarios and program structures to enhance and accelerate learning, planning, and decision-making. Wargaming in this context refers to a strategic simulation or exercise used by teams to anticipate and plan for potential challenges, opportunities, and outcomes in a complex environment. This capability would, notionally, integrate authoritative Department of Defense (DoD) data and knowledge sources with foundational generative AI models, utilize modern data caching and retrieval methods, and support flexible prompting strategies and workflows. The concept should consider how the application could scale to eventually exploit and integrate the range of relevant Defense acquisition data and knowledge sources such as is depicted in figure 1 below as a notional Defense Acquisition Body of Knowledge. Additionally, table 1 below provides a set of user stories to further elaborate on the capability DAU is seeking.

Figure 1 AI Driven Acquisition Decision Aid Application - Notional Architecture

The following user stories describe scenarios that the application/solution should support. This list is not exhaustive but is meant to elaborate on the functionality the solution should provide.

1 As a program manager, I need assistance in the drafting a tailored acquisition strategy

(AS) given a warfighting or business requirement, budget profile, and initial schedule goals.

2 As a program manager, I want to input my program's key parameters, such as budget, schedule, and performance requirements, so that the application can generate realistic scenarios and challenges that my team may face during program execution.

3 As a program team member, I want the application to generate potential threat adaptations, technology disruptions, and other future scenarios based on historical data and trends, so that we can develop contingency plans and risk mitigation strategies.

4 As a program team member, I want to collaborate with my teammates in a virtual wargaming environment generated by the application, so that we can simulate different scenarios and test our strategies in a risk-free setting.

Table 1 – Clafifying User Stories

D. Response Requirements:

Responses to this notice must include:

1. Company name, address and phone number

2. Contractor and Government Entity (CAGE) code and DUNS number. Indicate if your company is a large or small business under NAICs code 541715 Artificial intelligence research and development laboratories or services, NAICS code is 541511, Applications software programming services, custom computer, or other NAICs Code, and if it is registered as a vendor under a socio-economic program, such as 8(a), HubZone, Service-disabled veteran-owned, Small & disadvantaged, Woman-owned, etc.

3. Primary Point of Contact name, phone and email address

4. A summary description (limited to 3 pages) that demonstrates the ability to meet the aforementioned objectives.

5. If respondent has a government contract vehicle such as a Federal Supply Schedule or other government-wide use contract that would address the scope of this requirement, please indicate contract number in your response. Inform DAU if the vendor’s solution is on a pre-priced contract (GSA, NETCENTS, ESI, other) and provide contract number and URL / contracting point of contact information.

5 As a program manager, I want the application to continuously learn from our wargaming exercises and real-world experiences, so that it can provide increasingly accurate and relevant insights over time.

6 As a program manager, I want the application to facilitate cross-functional collaboration by engaging team members from different disciplines, such as engineering, finance, and logistics, in the wargaming process.

7 As a program team member, I want the application to provide me with personalized learning resources and recommendations based on the skills and knowledge gaps identified during the wargaming exercise, so that I can continuously improve my performance.

8 As a systems engineer, I want the application to suggest alternative system architectures and technologies based on the outcomes of the wargaming exercise, so that we can optimize our design for cost, schedule, and performance.

9 As a cost estimator, I want the application to generate realistic cost scenarios based on historical data and market trends, so that we can develop accurate cost estimates and budget projections.

10 As a financial manager, I want the application to suggest budget allocation and investment strategies based on the insights gained from the wargaming exercise, so that we can optimize our financial resources.

11 As a life-cycle logistician, I want the application to generate scenarios that test the resilience and effectiveness of our proposed supply chain and logistics network, so that we can identify potential bottlenecks and vulnerabilities.

12 As a contracting officer, I want the application to generate scenarios that simulate different contracting strategies and vendor behaviors, so that we can assess the risks and benefits of different procurement approaches.

13 As a cyber security professional, I want the application to simulate the potential consequences and impacts of a successful cyber breach on our acquisition program, so that we can develop effective incident response and continuity plans.

6. Questions Submission: Interested offerers must submit any questions concerning this RFI no later than 4:00pm (EDT), Thursday 27 June 2024. E-mail questions to Chris Robinson (Christopher.Robinson@dau.edu) and John Robey (John.Robey@dau.edu)

7. Submissions are limited to fifteen (15) pages (not including cover page). White paper submissions shall address the following elements:

a. Concept Description: Provide a detailed description of the proposed generative AI-based application, including its key features, functionalities, and user interface that address all aspects of the objective described above.

b. Technical Approach: Describe the technical approach for implementing the generative AI based application, including the proposed architecture, generative AI models, user interface, workflows (including agent-based workflows), data engineering and integration, computing platform, and IT infrastructure. Technical approach should favor use of existing DoD data, computing platform resources, and IT infrastructure.

c. Data Integration: Discuss how the application would integrate/interface with existing defense data sources and knowledge repository systems (see figure 1 for a notional list of relevant types and sources of data), ensuring data security and integrity. Address the need for the application to operate at Information Impact Level 5 (IL5) with the potential for Impact Level 6 (IL6) as defined in the DoD Cloud Computing Security Requirements Guide (CC SRG).

d. Application Optimization: Discuss cost effective approaches to generative AI model tuning, reinforcement learning, semantic data caching, and retrieval augmented generation (RAG) to improve overall performance of the application in terms of speed and accuracy.

e. Stakeholder Collaboration: Explain how the application will facilitate collaboration among various stakeholders involved in the defense acquisition process, such as program managers, technical experts, and decision-makers.

f. Decision Support: Describe how the application will provide decision support capabilities, such as scenario analysis, risk assessment, and trade-off analysis, to aid in the planning and decision-making process.

g. Scalability and Adaptability: Discuss the scalability and adaptability of the proposed solution to accommodate varying program sizes, complexities, and evolving requirements. Describe approaches for updating the AI system with advancements in technology.

h. Deployment and Maintenance: Provide an overview of the proposed deployment and maintenance strategy, including any required infrastructure, training, and support.

i. Intellectual Property: Clearly state any intellectual property rights associated with the proposed solution and any licensing or usage restrictions.

j. Security Measures: Compliance with Department of Defense (DoD) cybersecurity, Federal Risk and Authorization Management Program (FedRAMP), and National Institute for Standards & Technology (NIST) Risk Management Framework (RMF)/Authorization to Operate (ATO) processes.

k. AI Ethics and Compliance: The concept should address the provisions in Executive Order 13960 on Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government and the Chief Digital and Artificial Intelligence Office (CDAO) Responsible AI Toolkit.

l. Cost Considerations: While this market research is NOT a request for proposal/quote, to help determine affordability, the capability statement should include pricing models and a high level ROM for the described solution’s recurring and non-recurring costs to include installation /set-up, initial training, licensing type and terms (include copy or URL), number of users, roles, and how “user” is defined based on user roles, sustainment costs, upgrade costs, and other associated services. Please clarify/segregate specifically what is included in the estimated price and what would be an add-on cost. If your pricing is included on your website for the solution you describe in your capability statement, please provide the URL with any explanation of “included in described solution as shown in capability statement” or “would be an add-on cost” if not specifically shown on the website.

m. Vendor Experience and Capabilities: Describe previous experience in developing and deploying similar AI solutions, related research and development capabilities, and support and maintenance services offered.

n. Please be specific if your solution can meet the objectives or if other software add-ons are needed to meet the full objective. If add-on’s are needed, provide details on the add-on’s, the interfaces, and if your company provides the full solution or if DAU would need to acquire compatible functionality separately for integration. Also indicate if the objective can be met with no customization, minimal customization, marginal customization or extensive customization required.

8. Please provide contact information to arrange follow-on discussions with DAU. For Market Research purposes, the government may request an opportunity to connect for additional information.

E. Deadline:

The Government requests interested parties to submit a white paper and statement of current ability to supply the above capability via email with the subject line “DAU Generative AI Acquisition Planning and Decision Capability” to christopher.robinson@dau.edu, (and copy maria.t.evans.civ@mail.mil) by 2:00 pm (EDT), Monday, 29 July 2024.

F. Disclaimer and Important Notes:

This Request for Information is for market research purposes only and does not constitute a Request for Proposal/Quotation; and, it is not considered to be a commitment by the Government to award a contract nor will the Government pay for any information provided;

no basis for claim against the Government shall arise as a result from a response to this Request for Information or Government use of any information provided. Failure to submit information in sufficient detail may result in considering a firm as not a viable source and may influence competition and set-aside decisions. Regardless of the information obtained from this Request for Information, the Government reserves the right to consider any arrangement as deemed appropriate for this requirement. Respondents are advised that the Government be under no obligation to acknowledge receipt of the information received or provide feedback to respondents with respect to any information submitted.

No proprietary, classified, confidential, or sensitive information should be included in your response to this RFI Notice. The Government reserves the right to use any information provided by respondents for any purpose deemed necessary and legally appropriate, including using technical information provided by respondents in any resultant solicitation.

At this time, no solicitation exists; therefore, DO NOT REQUEST A COPY OF THE SOLICITATION and DO NOT SUBMIT OFFERS.

G. Points of Contact:

Chris Robinson, Associate Dean DAU Capital & Northeast Region Phone 240-925-2475 Email Christopher.robinson@dau.edu