RFI - Veterans Affairs AI platform 3rd party support services.docx

VA Enterprise Artificial Intelligence Support Services September 22, 2026

1. Purpose:

This is a Request for Information (RFI) only issued for conducting market research. Accordingly, this RFI constitutes neither a Request for Quote (RFQ), Request for Proposal (RFP), nor a guarantee that one will be issued by the Government in the future; furthermore, it does not commit the Government to contract for any services described herein. The Department of Veterans Affairs (VA) is not, at this time, seeking proposals or quotes, and therefore, will not accept, review, or evaluate unsolicited proposals or quotes received in response hereto. This notice is not to be construed as a commitment on the part of the Government to award a contract, nor does the Government intend to pay for any information submitted because of this request. The Government does not reimburse respondents for any costs associated with submission of the information being requested or reimburse expenses incurred for responses to this RFI. The information provided may be used by VA in developing its acquisition strategy and requirements documentation. Any information submitted by respondents to this RFI is strictly voluntary; however, any information received shall become the property of the Government and will not be returned to the respondent. Respondents are responsible for clearly marking any information they consider proprietary or commercially sensitive on a page-by-page basis. VA will use reasonable measures to protect legitimately designated proprietary information, but cannot guarantee confidentiality under applicable law, including the Freedom of Information Act (5 U.S.C. § 552). Interested parties are responsible for adequately marking proprietary, restricted, or competition sensitive information contained in their response.

2. Background:

Artificial Intelligence at Veterans Affairs VA serves 18.25 million Veterans and their families, administering benefits, delivering clinical care, and providing Veteran-facing digital services. The Veterans Health Administration, VA's clinical arm, is the largest healthcare system in the U.S., serving roughly 9 million Veterans annually across 1,200 facilities.

VA is committed to harnessing AI to deliver faster, higher-quality, more cost-effective services while upholding the highest standards of safety, security, and equity.

Five priorities:

1. Expand AI access across the workforce

2. Build an AI-ready workforce

3. Reimagine high-impact workflows with AI, automation, and agentic capabilities

4. Invest in data and infrastructure for AI adoption

5. Run transparent and effective AI governance

This is the third in a series of three Requests for Information (RFI) issued by VA in support of its enterprise AI acquisition strategy. The first two RFIs addressed the first-party AI tool described in Section 2 below, for context. This RFI addresses third-party services described in Section 3.

VA anticipates releasing the solicitation for third-party services early first quarter in Fiscal Year 2027 or calendar month October 2026. These dates are provided for industry planning purposes only, are subject to change, and do not constitute a commitment by the Government.

3. First-party AI product and services This is a separate acquisition described here for context on how it bounds this RFI's scope.

3.1 The product

VA intends to acquire through a separate acquisition an organization-managed, enterprise AI product. The product will provide conversational assistance, document and data analysis, enterprise knowledge retrieval, source-grounded research, business-document generation, coding assistance, and agentic task execution plus centralized administration of identity, users, groups, roles, models, connectors, agents, analytics, and monitoring.

3.2 First-party services

1. Account management and administrative onboarding/support (account director, customer success manager, engineer)

2. Adoption acceleration: use-case workshops, train-the-trainer/champion network support

3. Self-paced training curricula and adoption playbooks – AI foundations and per job-role

4. Safety, governance, and responsible-AI advisory aligned to federal regulations

5. Advisory collaboration on product roadmap inputs

6. Support developing outcome categories and metrics to measure impact on VA operations and Veterans

7. Applied AI product/design/engineering: architecture, connectors, reusable knowledge/tool packages, governed agents, evaluation harnesses, workflow integrations, and problem discovery

3.3 Surfaces and rollout

Surface will be web app only, desktop app only, or both. In year one, there may be a transition from one surface to two surfaces.

Surface decisions (which surface(s) are authorized and when) are made by VA Office of Information and Technology (OIT) under the first-party product contract's technology insertion process. Third-party services scope adjusts to support whatever surface(s) are authorized but does not independently drive product surface decisions.

Table 1. Rollout Schedule and Estimated Adoption Rates

MAU = Monthly Active Users. Active is defined as having sent at least one message.

4. Third-party services: What this Request for Information covers.

Immediately following the first-party acquisition, VA intends to procure third-party services. The third-party Contractor works in concert with VA and the first-party supplier(s). There is no incumbent.

Respondents should note that the relationship between the first-party product acquisition and this anticipated third-party services acquisition raises potential organizational conflict of interest considerations.

4.1 Objectives (VA Chief AI Office, Office of Information and Technology, year one)

4.2 Problem areas for third-party services requirement to address:

1. IT operations (see draft Performance Work Statement): Collaborate with VA teams on network, identity, device, endpoint, help desk, security, privacy, Authority to Operate (ATO), in-app administration, and sustainment. Areas to consider:

· Supported OS/browser/device/client versions

· Approved deployment, installation, config, update, and removal methods

· Network, certificate, proxy, firewall, identity, and admin dependencies

· Local data, file, telemetry, crash-log, and storage behavior

· Known compatibility limits and rollback procedures

· User support or digital service desk: users get reactions to issues as they arise

2. Discovery and adoption (customer success): Users achieve their desired outcomes while using a product. Help existing and new users successfully find and use the product, across both new and existing features and surfaces.

· At least one customer success lead per VA business unit, up to 50 total

3. Role-based virtual training: Default to problem-based learning: define the problem, success criteria, data, workflow redesign, and manage individual and group-level usage caps. Caps to be set by VA OIT with other business units. The Contractor has no authority to establish or modify usage caps.

Anticipated job roles for training (VA is open to industry feedback): medical/clinical, medical/research, analytics/data, IT operations, legal/compliance, software development, benefits/claims, education/training, human resources, contracting/acquisition, communications, finance/accounting.

4. Champion training: Deeper training for champion users who then train others.

5. Data connectivity: Integrate existing or new data connections; set VA-wide standard for connection setup and maintenance. VA defaults to open standards so connectors are reusable by any client.

6. VA-built AI assistants/agents: Small team to start, scaling with demand. See Section 5 for an example task addressing this problem.

7. Responsible AI adoption: Ensure the deployment and use of first-party product remains compliant with federal laws, regulations, and VA minimum guidance through communication, settings, and human oversight. The Chief AI Office may establish additional, product-specific requirements beyond VA’s, federal laws and regulation’s minimum guidance when greater product-specific safeguards are needed.

8. User feedback loop with VA and first-party product provider: capture user feedback to inform roadmap input.

9. Financial monitoring: regular ROI tracking and financial operations management.

10. Migration support: support VA and users through a product-to-product transition, if VA decides to migrate.

VA is open to consider additional problem areas industry identifies.

4.4 Scope boundary: first party vs. third party

Where scope is ambiguous, VA decides the problem priority and who is involved.

5. Example task: Priority applied AI development Illustrative task addressing Problem Area 7 (Section 4.2) VA-built AI assistants and agents.

Purpose: Standing development capacity to move prioritized AI use cases from validated concept to production-ready software. Exercised when demand exceeds baseline capacity, or a use case is ready to build. Self-contained. Award or exercise does not depend on any other task.

5.1 Scope

Full development lifecycle: discovery (requirements refinement), architecture, development, integration, testing, security/accessibility remediation, deployment, and post-deployment iteration, delivered in one of two modes:

5.2 Build-vs-buy decision gate

Before any use case enters build, the Contractor shall deliver a solution mode recommendation artifact for Government approval (COR + VA Chief AI Office), stating the recommended mode and rationale.

Build-vs-buy is a first-order design principle. The Contractor shall not duplicate capability a VA-authorized commercial product already provides natively (access control, Single Sign-On (SSO), observability, audit logging, usage metering, policy controls, agent scaffolding, connector frameworks). The recommendation artifact must state what native capability was evaluated and why custom development, if proposed, is necessary rather than duplicative. The recommendation artifact shall include a comparative capability matrix against the current native features of the VA-authorized commercial product(s), independent of the Contractor's own commercial relationships, and must disclose any commercial, contractual, or corporate relationship between the Contractor (or its subcontractors) and the product vendor being evaluated.

Deliverable:

A. Solution Mode Recommendation Artifact

5.3 Open standards and portability

The Contractor shall:

· Build integrations on Model Context Protocol (MCP) (or an equivalent open protocol) so a connector certified once works across VA surfaces and any MCP compatible tool

· Build reusable capabilities in the open Skills format (or equivalent) for portability across surfaces

· Instrument with OpenTelemetry (or equivalent) for monitoring, audit, and drift detection

· Design for model-agnosticism. Prompts, schemas, connectors, and configs must be demonstrably migratable between models and surfaces. Portability is an acceptance criterion, not a preference.

5.4 Safety and governance requirements

· Human in the loop at consequential steps

· Evidence-based results (pre-deployment evaluation, bias/fairness testing where rights-impacting, regression testing) required before any production promotion decision

· Section 508 conformance and applicable VA security requirements on all deliverables

· Connectors touching VA systems of record are read-only by default. Any write path requires separate approval, least-privilege scope, Government change control, and mandatory human approval for consequential actions

· Meet federal laws, regulations, and VA-specific policies and guidance

· Custom-developed code delivered as a full code package (source, docs, data models, schemas, architecture, config scripts) to the VA GitHub Source Code Inventory Repository throughout development, per the SHARE IT Act (P.L. 118-187, §6.1.13)

5.5 Deliverables

1. Solution mode recommendation artifact (per use case)

2. Technical design and architecture artifacts per approved build

3. Working software increments (connectors, assistants, agents, workflow integrations, and/or custom components, as applicable)

4. Custom-developed code package (if applicable), delivered continuously to VA repositories

5. Evaluation/test results supporting each production-promotion decision

6. Deployment and operations documentation (including observability and rollback)

7. Solution artifacts portfolio at task/increment close-out that note reuse potential

5.6 Acceptance criteria

Functional completeness against approved design and Definition of Done; demonstrated build-vs-buy discipline; conformance to open-standards/portability requirements; SHARE IT Act and data-rights compliance; Section 508 conformance; applicable VA security requirements.

INFORMATION TO INCLUDE IN A RESPONSE TO THIS NOTICE. Vendors responding to this notice are asked to provide the following:

1. Please describe your company’s capability to perform all required services in the Draft Performance Work Statement (PWS). As part of this response, please describe:

a. Any similar or previous experience supporting enterprise AI or generative-AI environments in the Federal Government or another regulated, high-security enterprise environment.

b. Identify relevant certifications, partner status, provider relationships, and direct escalation or reach back capabilities relied upon to perform the requirement.

c. Has the draft PWS provided sufficient detail to describe the technical requirements that are detailed in the PWS. If “NO”, please provide your technical comments/recommendations on elements of the draft PWS that may contribute to a more accurate proposal submission.

2. How much of this potential requirement would your company need to subcontract to other companies, if any?

3. Does the draft PWS contain sufficient information that would enable you to adequately respond to an RFP?

d. Identify material PWS requirements, assumptions, dependencies, risks, ambiguities, or barriers that could affect competition, pricing, staffing.

e. Is there enough detail to adequately price this requirement on a Firm-Fixed-Price basis? If not, what additional information would be required.

4. For priority areas 2-10, please describe representative services and deliverables that are provided in Industry or Government in these areas, and how these services are typically priced in the commercial marketplace. Please note - responses in this area will inform independent Government requirements development; however, any Offeror provided-language, task structure or deliverables will not be utilized in the Government’s final requirement.

5. Are there other considerations for these services that have not been covered, that the Government should be aware of when evaluating, selecting, and contracting for this support?

6. Technical Challenge Ideas: The Government intends to use a technical challenge as part of the evaluation of this requirement. What show-me-don’t-tell-me approaches should the Government consider for this procurement?

7. Firm Fixed Pricing and Payment Milestones: This requirement is expected to be priced as firm-fixed-price (FFP) for outcomes or deliverables tied, as much as practicable, to meeting the performance expectations and objectives defined in the PWS. The Contractor proposes a schedule of payment milestones to break the outcomes or deliverables into incremental payments. How would you recommend structuring payment milestones per period of performance? What basis for those milestones (for example, time-based, performance-based, deliverable-based, or a combination) would best balance predictable cash flow for the Contractor with meaningful accountability for the Government?

8. Rough Order of Magnitude (ROM) pricing for PWS Tasks 5.1 through 5.5 by task to include estimated labor types, labor mix and labor hours.

f. Identify the principal cost drivers and explain how pricing would change with user population, ticket volume, service hours, rollout waves, integrations, connections, agents or workflows, evaluation cycles, developer cohorts, use-case complexity, transition scope, or other measurable demand factors.

9. Interested Contractors shall also provide the following information in the submission:

· Name of Company:

· Address:

· Point of Contact:

· Phone Number:

· Email address:

· Company Business Size and Status:

· For VOSB and SDVOSBs, proof of verification in the Small Business Administration (SBA) VetCert Database.

· For firms who are not SBA VetCert verified, please provide VOSB and/or SDVOSB firms you are working with.

· NAICS code(s):

· System for Award Management (SAM) Unique Entity Identifier (UEI):

· Existing Contractual vehicles (General Services Administration (GSA)/ NASA Solutions for Enterprise-Wide Procurement (SEWP), etc.)

· The ability to comply with VAAR 852.219-75 - Notice of Limitations on Subcontracting—Certificate of Compliance for Services and Construction (Jan 2023) (Deviation)

2. Response Instructions:

Responses shall be submitted electronically by 10:00 AM Eastern Standard Time, October 7, 2026, via email to the Technology Acquisition Center (TAC) point of contact Contract Specialist Amber Davis, Amber.Quivers-Davis@va.gov and Contracting Officer, Joe Jones, Joseph.Jones6@va.gov. Please note “RFI# 36C10B26Q0834” in the subject line of your email response.

Responses shall not exceed 20 pages (excluding cover page and appendices) and shall be submitted in Microsoft Word format, with a minimum 11-point font and 1-inch margins. Responses exceeding the page limit may not be reviewed in their entirety. VA encourages concise, substantive responses over comprehensive but unfocused ones. NO MARKETING MATERIALS ARE ALLOWED AS PART OF THIS RFI. Generic capability statements will not be reviewed.

OFFERORS ARE ADVISED THAT THIS REQUIREMENT MAY BE DELAYED, CANCELED, OR REVISED AT ANY TIME