An official website of the United States government
A .mil website belongs to an official U.S. Department of Defense organization in the United States.
A lock (lock ) or https:// means you’ve safely connected to the .mil website. Share sensitive information only on official, secure websites.

A globe is shown with network icons (security, database & cloud), right aircraft carrier and ships.

There are currently one (1) Area of Interest open for solicitation

If you would like to be notified when new solicitations are posted, please fill out our interest form here.

For more information on responding to a solicitation, see our NAVWAR CSO Resource guide.




Area of Interest: Generative Artificial Intelligence (GenAI) Electronic Performance
Support System (EPSS)
 
Responses Due: 17-SEP-2026 at 2:00 PM PDT
Questions Due: 28-AUG-2026 at 2:00 PM PDT

The Government may not address all questions received. Questions and Answers will be posted
on the NAVWAR CSO Website: https://www.navwar.navy.mil/Work-With-Us/Open-CSOCalls/

Contact information for questions and technical/submission issues:
Jennifer.L.Zornow.civ@us.navy.mil

The Government will not pay respondents for the costs associated with the submission of white
paper/solution briefs.

Introduction:
PMW 160 is the Department of the Navy’s premier program office for tactical networks, which
delivers mission-effective, affordable, and cyber-resilient network capabilities to Navy tactical
forces operating globally. Naval Information Warfare Systems Command (NAVWAR) on behalf
of PMW 160 is seeking commercial solutions for a Generative Artificial Intelligence (AI)-
powered Electronic Performance Support System (EPSS).

Problem Statement:
Traditional training methods cannot scale fast enough to match the speed of software updates,
leading to a significant lag in the rapid adoption of newly deployed technologies by fleet
personnel. This disconnect creates critical skill gaps, imposes unsustainable cognitive workloads,
and inflicts chronic information overload on warfighters.

To keep pace with the rapid, agile development of modern tactical networks and the advanced
capabilities hosted thereon, the Department of the Navy is modernizing its Manpower, Personnel,
and Training (MP&T) and performance support frameworks. As system complexity increases,
PMW 160 is proactively seeking solutions to streamline cognitive workloads, mitigate
information overload, and accelerate technology adoption. Ensuring warfighters are equipped
with the advanced digital tools necessary to rapidly master and maintain modern network
capabilities is vital to sustaining long-term operational readiness and mission success.

To achieve these modernization objectives, the Navy requires a transformative, Generative AIpowered
EPSS along with comprehensive operations and sustainment support. The required
solution is not merely a software tool, but a strategic performance-enhancement framework that
integrates dynamic enterprise knowledge bases, interactive guides, and communication resources
to provide immediate, on-the-job assistance which, in turn, facilitates adoption of new
technology.

The primary goal of the GenAI-powered EPSS is to enhance warfighter performance by
delivering individualized, contextual guidance precisely at the moment of need. A GenAI powered EPSS acts as a "digital coach" embedded directly within the daily workflow,
empowering warfighters with self-help capabilities to perform at the speed of relevance and
maximizing the combat effectiveness of the modern Navy.

Primary Attributes:
• Strict Retrieval-Augmented Generation (RAG) Grounding & Explainable AI (XAI)
AI generation must be confined exclusively to approved organizational data. The system
must apply semantic filters to block hallucinations and provide verifiable citations for
every response.

• Hybrid RAG - The system must implement a hybrid retrieval architecture that
merges dense vector semantic search with sparse keyword/lexical search in parallel. The
retrieval engine must utilize a re-ranking algorithm to synthesize, deduplicate, and
prioritize the most contextually and nomenclature accurate data chunks (including
technical system terminology and organizational nomenclature) prior to Large Language
Model (LLM) generation.

• Zero Trust & Offline Security - The system must enforce Zero Trust architecture,
encrypting data at rest/transit via Federal Information Processing Standards (FIPS)-
validated cryptography. The system must verify identity locally using signed tokens when
disconnected.

• Access Control & Data Masking - The system must dynamically mask Personally
Identifiable Information (PII)/Protected Health Information (PHI) and enforce strict Role-
Based Access Control (RBAC), ensuring AI only synthesizes answers from data the user
is authorized to view.

• Graceful Degradation - The system must degrade gracefully to basic keyword search if
AI components fail and utilize automated watchdogs to self-heal/restart services without
bricking the workstation.

• Audit & Telemetry Logging - The system must archive comprehensive, tamper-evident
logs of user prompts, generated outputs, and source documents to support compliance,
legal, or financial reviews.

• Adversarial AI Defense (Prompt Injection) - The system must implement robust,
deterministic guardrails to protect the model against both direct and indirect prompt
injection attacks. The system must ensure the AI cannot be hijacked, bypassed, or
manipulated into executing unauthorized commands or accessing unauthorized data.

• Memory Optimization & Thrashing Prevention - Memory optimization guardrails
(e.g., dynamic context window management, automated offloading) must be implemented
at the runtime level to prevent system thrashing. The system must guarantee stable
execution without monopolizing host memory or freezing the operating system on
resource-constrained devices.

• Hardware Agnosticism & Central Processing Unit (CPU) Fallback - The inference
engine must dynamically prioritize Graphics Processing Unit (GPU) utilization when
available to maximize generation speed. However, the architecture must be strictly GPUagnostic
and must seamlessly fall back to CPU-only inference if a discrete GPU is
unavailable or fails.

• Technical Maturity and Demonstration Readiness – Vendor solution must currently
exist at a Technology Readiness Level (TRL) of 6 or higher (prototype system tested in a
relevant environment).

• Information Assurance & Common Vulnerabilities and Exposures (CVE)
Remediation Service Level Agreement (SLA) - Vendor must provide a binding SLA to
patch critical and high-severity vulnerabilities (CVEs) in the software, container images,
or models within 14 days of discovery to maintain Authority to Operate (ATO).

• "Air-Gapped" Update Packaging - Vendor must package all software updates, security
patches, and updated RAG vector databases into fully self-contained, offline install files
suitable for isolated, stand-alone network architectures.

• Active RAG Pipeline Support - Vendor must provide continuous support to ingest,
chunk, and index new organizational documents into the vector database without system
downtime.

• Critical Incident SLAs - Vendor must guarantee response times for Severity 1
(Critical/Catastrophic), such as System Down/Dangerous Hallucinations, and Severity 2
(Major/High) incidents, including rapid workarounds or rollbacks.

Desired Attributes:

• Hardware-Constrained Inference - The local software, models, and vector stores
should occupy < 8 Gigabytes (GB) of storage and run full generative inference utilizing <
6 GB of Video Random Access Memory (VRAM)/Random Access Memory (RAM) on
standard Commercial-Off-The-Shelf (COTS) CPUs (< 30 Watt draw).

• Conversational Orchestration - The system should support cross-app execution,
allowing users to naturally prompt the AI to pre-fill forms or move data across disjointed
enterprise software via Application Programming Interface (API) orchestration.

• Adaptive Scaffolding & Ergonomics - The User Interface (UI) should dynamically
adjust instruction complexity based on user performance, injecting "cognitive nudges" or
rotating tasks to prevent user skill degradation over time.

• Modular & Resilient Updates - Architecture should cleanly separate the core app, AI
model, and knowledge base to allow independent, resume-capable "delta-sync" updates
over intermittent network connections.

• Continuous Fine-Tuning Services - Vendor provides periodic instruction and prompt
fine-tuning to resolve newly discovered hallucination vectors, user-reported biases, or
changes in DoD operational terminology.

• Bandwidth-Aware Diagnostics - Technical support teams should be capable of
diagnosing system anomalies remotely via low-bandwidth "pull" diagnostic logs, without
requiring live screen-shares.

• Model Version Swapping - Vendor provides services to safely test, upgrade, and swap
the underlying LLMs (e.g., to a newer, more efficient open-source model) without
rewriting the core application.

• "Train-the-Trainer" & SME enablement - Vendor provides dedicated training for
internal Subject Matter Experts (SMEs) on how to manage, curate, and audit the AI's
knowledge base independently.

Additional Notes:

Cybersecurity and Information Assurance - Best practices must be implemented and strictly
adhered to. Measures must also be taken to protect the supply chain from malicious and/or
counterfeit software and hardware components to ensure safe, secure, and reliable operation of
the system.

Security - Companies must be able to meet personnel security clearance requirements as
designated by the Government.

Capability Pitches - Companies are expected to describe how their solution will meet
Government requirements via a presentation in an unclassified environment as part of Phase 2 of
the CSO AOI. Presentations must comprehensively address the following areas:

Technical Viability: Explain exactly how your commercial technology solves the
Government’s specific mission problem.

Rough Order of Magnitude (ROM): Provide a high-level, preliminary cost estimate and
commercial pricing model.

Project Schedule: Present a notional timeline and key milestones for the prototyping
phase.

• Data Rights and Intellectual Property (IP): Explicitly state your company’s position on
intellectual property (IP), technical data rights, and software licensing. Note any
commercial license agreements (End-User License Agreements/Service-Level
Agreements) that would apply to the Government’s use of the technology.

Live Q and A / Demonstration: Prepare your technical team to participate in a structured
question and answer session. Presenters must be ready to address specific risks, answer
evaluator questions, and explain how they identify, mitigate, and resolve potential
challenges.

To achieve specific mission objectives, the Government reserves the right to team companies
with complementary solutions under this AOI or combine capabilities across other NAVWAR
AOIs. Consequently, selected vendors may be required to partner with other awardees to deliver
a comprehensive solution.

Awarding Instrument

This solicitation will be awarded in accordance with the Commercial Solutions Opening (CSO)
process detailed within N00039-25-S-0001 (NAVWAR CSO).

Please refer to NAVWAR CSO for additional information regarding the three-phase process:
• Phase 1: Solution Briefs
• Phase 2: Pitch Session
• Phase 3: Proposal

Vendors are reminded that in order to utilize an Other Transaction (OT) agreement the
requirements of 10 USC 4022 must be satisfied. Specifically reference 10 USC 4022(d), which
requires significant contribution from a nontraditional defense contractor, all participants to be
small business concerns, or at least one third of the total cost of the prototype project is to be
paid out of funds provided by sources other than the federal government. Any resulting
agreement from this solicitation will include language requiring your company to confirm
compliance with Section 889 of the John S. McCain National Defense Authorization Act for
Fiscal Year 2019 (Pub. L. 115-232). If you are unable to confirm compliance with the referenced
law, the Government will not be able to enter into an agreement with your company.

Potential Follow-On Contract or Agreements

Companies are advised that any Other Transaction (OT) agreement for Prototype awarded in
response to this solicitation may result in the direct award of a follow-on production contract or
agreement without the use of further competitive procedures.

The follow-on production contract or agreement will be available for use by one or more
organizations within the Department of Defense (DoD). As a result, the magnitude of the followon
production contract or agreement could be significantly larger than that of the Prototype OT
agreement. All Prototype OT agreements will include the following statement relative to the
potential for follow-on production: “In accordance with 10 U.S.C. § 4022(f), and upon a
determination that the prototype project for this transaction has successfully been completed, this
competitively awarded Prototype OT agreement may result in the award of a follow-on
production contract or transaction without the use of competitive procedures.

Response Instructions

Per the N00039-25-S-0001 NAVWAR CSO announcement, vendors interested in responding to
this AOI are invited to submit Phase 1 Solution Briefs. Solution Briefs should not exceed five (5)
written pages using 12-point font or, alternatively, Solution Briefs may take the form of briefing
slides which should not exceed fifteen (15). These limits are not requirements but are strongly
recommended.

Title Page (does not count against page limit)

Company Name, Title, Date, Point of Contact Name, E-Mail Address, Phone, and Address.
Specifically identify the AOI for which the Solution Brief is submitted.

Executive Summary (one page limit, included in five (5) page limit)

Provide an executive summary of the technology.

Technology Concept

Describe the unique aspects of your technology and the proposed work as it relates to the
AOI. Identify whether the effort includes the pilot or demonstration of existing commercial
technology (identified as commercially ready and viable technology), or the development of
technology for potential defense application. If development or adaptation is proposed,
identify a suggested path to mature the technology. Identify aspects which may be considered
proprietary.

Company Viability

Provide a brief overview of the Company. Provide a summary of current fundraising to date
or a summary of the top line (gross sales/revenues). Provide a summary of product
commercialization and go-to-market strategy.

Recommended Award Type

Companies should recommend an award type to be considered for the effort (FAR-based
Procurement Contract or OTA). Should the Company be asked to submit a Proposal, the
Government and Company will further negotiate the best type of award for the effort.

Submission Instructions
Vendors interested in submitting a response to this AOI must submit their solution brief via the
Procurement Integrated Enterprise Environment (PIEE) Solicitation module link provided in this
posting. The Government will not accept emailed submissions. If you have any questions
regarding submission, you may contact the Government point of contact identified in the
SAM.gov posting/notice.

PIEE Public Notice:
https://piee.eb.mil/sol/xhtml/unauth/search/oppMgmtLink.xhtml?noticeId=N0003926R9450&no
ticeType=SolicitationNotice


SAM.gov Public Notice:
https://sam.gov/workspace/contract/opp/61a92797c0634d7c89266cb5573675f3/view


GenAI EPSS Questions and Answers, Set 01

PAO Release Number: NAVWAR HQCA-2026-A-085
 
  1. For this effort, would the government be able to provide the current training guidelines/documents used for the current (what sounds like manual) processes? If not, could the government provide a fake document that has the general outline of what these training documents look like? Ample open-source data regarding Government training should be available via searchable means. Responses should focus on capability and services requested, not on current Department of Navy (DoN) training guidelines, processes, or data.
 
  1. Could the Government share any additional details and next steps or advice? The goal of the AoI is to seek innovative commercial products and/or solutions from industry on capability and services requested. Please refer to the NAVWAR CSO announcement for the next phases.
 
  1. The AoI directs Phase 1 Solution Briefs to the PIEE Solicitation Module. CSO N00039-25-S-0001 states that companies responding acknowledge they must be registered in SAM prior to award. PIEE vendor registration requires a CAGE code, and PIEE validates that code against SAM. Could the Government confirm whether Offerors may submit a Solution Brief while CAGE/SAM registration is pending? All Offerors need active SAM registrations at the time of submission.
 
  1. Will the Government review Phase 1 Solution Briefs that are at a lower TRL level? The Government requires a TRL 6 solution. However, the Government will review all white papers submitted to this AoI.
 
  1. Will the Government offer meetings with Offerors during this stage? At this time, the Government is seeking solution briefs under Phase I of the CSO. Once the Government has evaluated those vendor solutions to be of merit, they will be invited to provide a pitch under Phase 2, but not before.
 
  1. Could the Government please clarify the acceptable scope of a "commercial solution?” Is the Government seeking exclusively Commercial Off-The-Shelf (COTS) software products, or does a mature Government Off-The-Shelf (GOTS) software capability with demonstrated success at Technology Readiness Level (TRL) 6 or higher satisfy the definition of a commercial solution for this requirement? The Government requires a TRL 6 solution. Per the AOI, the software must run on standard COTS CPUs.

8. Which specific warfighter tasks, mission workflows, and tactical-network use cases should be prioritized for the prototype? 

The Government is not seeking a solution that is specific to a Warfighter task, workflow, or use case. The Government is seeking a solution capable of addressing the primary and desired attributes in the AoI for various warfighting scenarios. 

 

9. Which user roles, organizations, experience levels, and operating locations are expected to use the EPSS during the prototype? 

The Government anticipates that the selected solution will support System Administrators, Operators, and maintainers roles, experience levels, and operating locations afloat and ashore. 

 

10. Can the Government provide additional details on: 

- How many users, concurrent users, workstations, and deployment locations should the prototype and Rough Order of Magnitude estimate assume? 

- What response-time and performance thresholds must be supported? 

- Which categories of PII and PHI will be stored? 

- What are the data auditing requirements? 

- Is there a prescribed target hardware and deployment architecture? 

- Will CUI be provided to Offerors and what are the handling requirements? 

- What user roles and authorization levels must be implemented? 

- What test data, scenarios, and acceptance criteria will be used? 

- What are the requirements for graceful degradation and recovery? 

- What system/data security requirements must be met? 

Initial assessment will focus on primary and desired attributes as defined in the AoI. The Government will provide additional details to Offerors who are selected to proceed to Phase 2, Pitches. 

 

11. What operating systems, workstation types, CPU/GPU configurations, and other target hardware baselines must the solution support? Are the stated limits of less than 8 GB storage, less than 6 GB VRAM/RAM, and less than 30-watt CPU draw mandatory thresholds or desired objectives? 

The specifications defined in the AoI are industry standards and should be treated as minimum compute and store requirements. 

 

12. What organizational data sources are approved for grounding, including expected document types, formats, data volumes, languages, classification levels, and update frequency? 

The Government anticipates that data sources will include technical documentation, multi-media, and software-based tools or repositories with file types including, but not limited, to MS Office suite files, Adobe products, and standard video file formats. 

 

13. Will the Government provide the initial document corpus, metadata, technical terminology, and organizational nomenclature needed to build and evaluate the RAG knowledge base? 

The Government will provide information and artifacts to those who are selected for prototyping. 

 

14. How should the system resolve conflicting, obsolete, duplicate, or incomplete source documents, and will the Government identify which sources are authoritative? 

The Government will provide only authoritative documentation and information and provide updated artifacts, as needed. 

 

15. Who will approve new content for ingestion, and what review, release, and update process should the active RAG pipeline support? 

The Government will provide new content as required. 

 

16. What citation format and level of source granularity are required for generated responses, and what should the system display when no authorized, sufficiently grounded answer is available? 

The goal of the AoI is to seek innovative commercial products and/or solutions from industry on capability and services requested. The Government will not provide next steps or advice. 

 

17. What are the required Severity 1 and Severity 2 response, workaround, restoration, and rollback times; required support hours; low-bandwidth diagnostic constraints; and process for delivering the required 14-day CVE remediation packages to air-gapped systems? 

See Government response to question number 0010. 

 

18. What prototype duration, milestones, demonstration environment, Government-provided data or equipment, success criteria, pricing period, licensing assumptions, data-rights expectations, and anticipated teaming or integration requirements should vendors use for the Phase 1 solution brief and Phase 2 pitch? 

Phase 1 should focus on the primary and desired attributes outlined in the AOI letter. The Government will provide additional details to Offerors who are selected to proceed to Phase 2, Pitches. 

 

19. Is this requirement related to the MyNavy HR modernization effort?  

No, this is not related to the MyNavy HR effort. 

 

20. What is the anticipated funding range for the initial GenAI EPSS effort?  

At Phase 2, selected vendors are expected to provide a preliminary cost estimate and commercial pricing model. 

 

21. Hardware-Constrained Inference:  

- Specs are given for HW. Is this existing HW that the SW Solution will be deployed to or can new HW be considered.  

- If existing system, are the included HW constraints the only data available and will it vary per install  

The Government will consider alternative hardware solutions. Specifications and/or constraints will vary by platform type. Note: The specifications defined in the AoI are industry standards and should be treated as minimum compute and store requirements. 

 

22. Does the active SAM registration requirement apply only to the entity submitting the Phase 1 solution brief as the offeror or prime contractor, or must every named subcontractor and teaming member also have active SAM registration? 

All Offerors including subcontractors need active SAM registrations at the time of submission. 

 

23. Will Solution Briefs be evaluated on the Primary Attributes alone, or will the Desired Attributes also be evaluated? If both, is there a relative weighting between them? 

The Government will evaluate on both Primary and Desired attributes. Primary attributes carry higher weighting. 

 

23. Is the capability described in this AOI intended to operate independently, or to interoperate with existing Department enterprise artificial intelligence services? 

The Government anticipates eventual interoperation with other DoD Solutions. 

 

24. Is this intended to replace an existing EPSS or augment current training and support systems? 

The GenAI EPSS solution will augment current training and support systems. 

 

25. Beyond ingesting existing documentation, does the Government anticipate a need to restructure or author content for point-of-need use (e.g., decomposing manuals into procedural, decision-point, and troubleshooting formats)? Is that content engineering expected to be vendor-provided? 

Yes, the Government will require Vendors to engineer content. 

 

26. Will the Government support access to representative fleet personnel for evaluation during the prototype phase?  

Yes, Government personnel will be involved in evaluation during Phase 3, Prototyping. 

 

27. Post-fielding, how does the Government intend to monitor whether the EPSS is achieving reduced cognitive workload and accelerated technology adoption? Is continuous effectiveness evaluation anticipated within operations and sustainment? 

The Government anticipates monitoring this during Phase 3, Prototyping. 

 

28. How does the Government scope "relevant environment" for this AOI: does a prototype demonstrated in another Government or federal operational environment satisfy the attribute, or must the solution have been tested on a Navy tactical network or an equivalent afloat configuration? Additionally, is the TRL 6 determination made against the integrated solution as a whole, or attribute by attribute? 

The TRL 6 will be evaluated as a whole. A relevant environment includes environments outside of a lab environment. 

 

29. The Area of Interest states that award will be made in accordance with the CSO process detailed within N00039-25-S-0001 and refers respondents to that announcement for the three-phase process and response instructions. Would the Government confirm where the current version of N00039-25-S-0001, and any amendments, may be obtained? 

N00039-25-S-0001 is available on SAM.gov. 

 
Guidance-Card-Icon Dept-Exclusive-Card-Icon