HomeAgenciesVA AI Use Case Inventory Counts 367 Systems. The Near-Term Move Is...

VA AI Use Case Inventory Counts 367 Systems. The Near-Term Move Is Two Open IDIQs, Not the RFI

VA disclosed 367 AI use cases, the third most of any federal agency behind only HHS and NASA. I checked what's actually open to compete for today: a closed RFI, two open IDIQs, and budget lines Congress hasn't enacted.

367 Disclosed AI Systems, Not One Disclosed Contract Vehicle

VA disclosed 367 AI use cases in 2025, the third most of any federal agency behind only HHS and NASA. That count reads like an open AI market you can’t bid into. The VA AI use case inventory is a disclosure document, not a pipeline. Today’s real work runs through two already-open IDIQs.

VA splits the 367 into 13 “Consolidated” systems, its shared enterprise infrastructure, and 354 narrower tools built for a single clinic or workflow. No disclosed vehicle sits behind any of them. What moves your pipeline: two already-open IDIQs, T4NG2 and SPRUCE, plus a since-closed RFI that would have been the clearer path to a new VA-wide award. If you’re an AI product firm, systems integrator, or SDVOSB without a position on either vehicle, that’s the gap worth checking. Treat it as a signal to watch, not a live bid, until one specific thing changes.

Three facts are true at once. Most coverage of AI in government collapses them into one number:

  • VA’s 367/215 count is a disclosure requirement under OMB Memorandum M-25-21, not a new contract action.
  • An RFI for enterprise AI infrastructure, which named no incumbent, closed its response window June 9, 2026. No follow-on solicitation has been posted as of this writing.
  • Two already-open VA IDIQs, T4NG2 and SPRUCE, already cover AI-adjacent scope and don’t require VA to run a new enterprise-wide competition first.

What follows: those facts, requested versus enacted money, and what I’d check before your BD lead and capture lead treat this as more than a signal.

What the VA AI Use Case Inventory Actually Counts

The VA AI Use Case Inventory lists 367 AI use cases, spanning clinical care, benefits processing, records management, and internal operations. VA last updated it in December 2025 and refreshed its risk-management fields in April 2026. Of the 367, 215 are classified “high-impact,” the OMB risk tier for AI systems that materially affect rights, safety, or benefits decisions.

This isn’t a new kind of disclosure. VA has filed a comparable inventory since at least OMB Memorandum M-24-10 in 2024. That memo built on a 2020 executive order requiring federal AI use-case inventories. M-24-10 already required minimum risk-management practices for “rights-impacting” and “safety-impacting” AI, on a compliance clock that ran out in December 2024. What’s new: OMB rescinded and replaced M-24-10 with M-25-21 on April 3, 2025. That folded those categories into a single “high-impact” tier and reset the clock. Agencies got 365 days from issuance. Every high-impact use case had to meet minimum risk-management practices by April 3, 2026, or carry a documented waiver.

The 354-System Gap Between “Disclosed” and “Open”

Only 13 of VA’s 367 use cases appear in a separate “Consolidated” inventory. That’s VA’s own label for tools adopted broadly enough to work as shared enterprise infrastructure. Three named examples:

  • Ambient AI Scribe listens to clinician-patient conversations and drafts clinical notes.
  • STORM is a clinical decision-support tool for opioid risk. VA’s own Health Services Research and Development office credits it with a 22% reduction in mortality among the high-risk veterans it monitors.
  • A payment-redirect-fraud model flags suspicious direct-deposit changes before a payment goes out, catching an estimated one to two fraudulent redirects per 1,000, per VA News.

The other 354 use cases are narrower point solutions VA built for a specific clinic or workflow. That gap doesn’t mean VA is overstating its AI adoption. VA discloses both figures itself. And 354 pilots on top of 13 enterprise systems is a normal shape for a multi-year AI ramp. The gap does mean this: treating “367 AI systems” as a market-sizing number overstates what you can compete for. VA’s public inventory doesn’t say whether any of the 354 carries recompetable follow-on work.

VA reported 215 high-impact AI use cases in 2025, the most of any federal agency, out of 445 high-impact cases OMB counted government-wide. VA’s total, 367, ranks third government-wide, behind only HHS (447) and NASA (425). (VA AI Use Case Inventory, department.va.gov, December 2025; government-wide figures per Nextgov/FCW and Global Government Forum reporting on OMB’s 2025 Federal Agency AI Use Case Inventory, 2026-04)
“We want to make sure that we’re pushing the envelope from an AI-enabled to an AI-first environment.” (Howard Steinman, Senior Advisor, VA Office of Information and Technology, per GovCIO Media & Research)

The Enterprise AI RFI VA Posted, and Why Its Window Already Closed

VA OIT, under Chief AI Officer and CTO Charles Worthington, posted a request for information on May 29, 2026. It’s titled “VA Enterprise Artificial Intelligence,” notice VA-26-00070193. Responses were due June 9, 2026, eleven days later. The RFI asks industry for two things: front-end AI infrastructure (the user-facing interface) and back-end infrastructure (API access to large language models). VA wants both to expand assistive, collaborative, and agentic AI capabilities workforce-wide. Per FedScoop’s reporting, VA said it is “actively piloting tools from multiple Frontier AI Labs to evaluate user experience, performance, and value at enterprise scale.”

FedScoop and OrangeSlices AI both report that the effort has no incumbent contractor. I couldn’t confirm that against SAM.gov’s own posting text. The page didn’t render for direct review this pass. Treat the no-incumbent detail as sourced to secondary reporting, not a primary-document read. The window closed with no follow-on posted, so you can’t respond to this RFI today. What it does tell you: VA OIT is shaping a future acquisition for shared AI infrastructure. Whether that connects to the Decision Intelligence and Automation budget line below is my read. Neither document says so. A formal solicitation, if one comes, would move this from a signal to something you can act on.

Two Open Acquisition Vehicles Can Carry This Work Without a New VA-Wide Award

You don’t have to wait on VA’s next enterprise AI solicitation. Two existing IDIQ vehicles already cover scope that AI integration work fits inside. One of them is reserved entirely for service-disabled veteran-owned small businesses.

Vehicle Ceiling Set-aside Relevant scope Status
T4NG2 $60.7B Full and open, roughly 17 of 33 primes SDVOSB Systems engineering, data analytics, health IT Cleared to issue task orders, March 2026
SPRUCE $2.4B 100% SDVOSB Software and digital services Open, awarded
VA-26-00070193 (RFI) Not stated N/A, market research Enterprise AI infrastructure Response window closed, no follow-on yet

T4NG2 cleared to begin issuing task orders in March 2026. A Court of Federal Claims ruling had resolved a multi-year bid protest, per Truvisory’s reporting. The ceiling and prime-count figures above come from trade press, Market Connections and Truvisory, not from USASpending or FPDS. Treat the exact numbers as approximate. Neither vehicle’s listing states an on-ramp window or an ordering-period end date. I can’t tell you how much runway either has left.

What’s solid: both vehicles are open and awarded today. Whether a given piece of AI work falls inside either vehicle’s scope is a contracting officer’s call. I’m reading scope off Tier 3 summaries, not either IDIQ’s statement of work. A task order under an already-awarded IDIQ doesn’t require VA to run a new department-wide competition first. Mid-market firms already positioned on either vehicle can pursue that work now.

Real Money Versus Requested Money in VA’s FY2027 AI Budget

None of VA’s FY2027 AI money is enacted. Both lines below are requests in the President’s Budget. The two outlets that reported them expect Congress to change the figures during markup.

Here’s what VA’s FY2027 budget submission requests for AI:

  • $47.8 million requested for the Decision Intelligence and Automation account, a $4.7 million, 10.9% increase over FY2026 enacted.
  • $130 million requested for VBA claims-processing automation and AI, a separate line.
  • Combined, that’s $177.8 million in FY2027 AI-specific requests, sourced to VA’s own Budget Submission and corroborated identically by Nextgov and Government Executive.

Neither figure is available money until Congress acts. The same Budget Submission attributes the Decision Intelligence and Automation increase to an “AI Infrastructure solution.” That subject matter overlaps what the RFI describes. But the budget line’s name isn’t the RFI’s title, and no document links the two. That overlap is why I infer a follow-on acquisition action is coming. It isn’t evidence one exists.

Is This PURSUE, MONITOR, or DEFEND for Your Pipeline?

My answer: MONITOR, not PURSUE. DEFEND doesn’t apply either, since T4NG2 and SPRUCE are open multi-award IDIQs with no single incumbent VA is displacing. The RFI that was the clearest path to a new VA-wide award has closed. No AI-specific task order is confirmed under either vehicle. The FY2027 budget lines are requests, not appropriations. That’s a real signal, not yet a live one. Here’s how I’d work it while it’s still a MONITOR:

  1. Check whether your company holds a prime or subcontract position on T4NG2 or SPRUCE. If it does, pursue task-order work there now.
  2. If you don’t hold a position, evaluate teaming with a current prime. I can’t point you to an on-ramp date, but I’d start those conversations now.
  3. Set a SAM.gov watch for any notice referencing VA-26-00070193 or “VA Enterprise Artificial Intelligence.” A posted follow-on, or a confirmed AI-specific task order, moves this from MONITOR to PURSUE.
Pull the SPRUCE and T4NG2 awardee lists from USASpending or SAM.gov today. If you hold a position on either vehicle, check whether your account team has raised VA OIT’s AI infrastructure needs with your government customers yet. If you don’t, closing that gap, not chasing the 367-use-case number, is what your BD lead should work next.

What I Couldn’t Verify, and What Would Change This Read

VA’s own compliance plan confirms it met the April 3, 2026 M-25-21 compliance date. FedScoop’s reporting on OMB’s tracking puts VA alongside Labor, NASA, State, GSA, and EPA. What I couldn’t confirm: whether any of VA’s 215 high-impact use cases received a waiver rather than clearing the bar outright. OMB requires agencies to publish waiver summaries. VA’s plan doesn’t list any, and I found no separate disclosure.

  • Verify SAM.gov’s own posting text for VA-26-00070193 confirms no incumbent contractor.
  • Confirm T4NG2’s and SPRUCE’s ceiling and prime-count figures against USASpending or FPDS.
  • Ask your VA program contacts whether any of VA’s 215 high-impact AI use cases carry an OMB M-25-21 waiver.

FAQ

What did VA’s AI inventory report actually find?

VA disclosed 367 AI use cases in its 2025 AI Use Case Inventory. Of those, 215 are classified high-impact, the most of any federal agency, out of 445 high-impact cases OMB counted government-wide. VA’s total count of 367 ranks third government-wide, behind only HHS and NASA. Only 13 of the 367 appear in VA’s separate “Consolidated” inventory of widely adopted, enterprise-scale systems.

Is VA’s Enterprise AI RFI, VA-26-00070193, still open?

No. VA posted the RFI on May 29, 2026, with responses due June 9, 2026. That window has closed, and I found no follow-on solicitation posted as of this writing. Watch SAM.gov for a formal acquisition action referencing this notice.

How much of VA’s AI spending is real, enacted money right now?

Very little that’s AI-specific. The $47.8 million Decision Intelligence and Automation line and the $130 million VBA claims-automation line are both FY2027 budget requests, not enacted. Nextgov and Government Executive both expect Congress to change the figures in markup.

Which VA contract vehicles can carry AI work today without a new award?

Two: T4NG2 and SPRUCE. T4NG2 is a $60.7 billion ceiling IDIQ, not obligated spend, cleared to issue task orders since March 2026. SPRUCE is a $2.4 billion ceiling IDIQ, also not obligated spend, set aside entirely for service-disabled veteran-owned small businesses. Both already cover systems-engineering, data-analytics, or software scope that AI integration work fits inside.

If I’m not positioned on T4NG2 or SPRUCE, does this VA AI inventory actually change my pipeline?

Not immediately. The 367-use-case count isn’t a bid opportunity, and the one RFI that was a near-term signal has closed. Your realistic options: team under an existing T4NG2 or SPRUCE prime, or watch for a follow-on solicitation to VA-26-00070193. Treat this as a MONITOR signal you can convert to PURSUE the moment either event happens.

BD classification: MONITOR Confidence: Moderate Evidence that would upgrade this to PURSUE: A follow-on solicitation posted superseding VA-26-00070193, or a confirmed AI-specific task order issued under T4NG2 or SPRUCE. Evidence that would downgrade this: VA cancels or no-bids the enterprise AI effort with no follow-on action, or the FY2027 AI budget lines are zeroed out in enacted appropriations.

The Contract Opportunity Atlas

Two issues a week.. Free.

Two issues a week. Data-driven intelligence for small tech firms selling to the federal government. Free.

Subscribe to Contract Opportunity Atlas

Get federal technology, AI, procurement, and GovCon insights delivered to your inbox.

Shahid Shah
Shahid Shah
Shahid specializes in bringing world-class CTO, CISO, and EiR expertise to startups, business units and companies on a part-time (fractional) basis. With a rich background in regulated, safety-critical industries like Med Devices, Digital Health, and Gov 2.0, he possess a unique understanding of complex, high-demand products and services. He is a C-suite native that can easily blend in with technical and engineering teams that need to deliver revenue-generating solutions to the marketplace. He has served as an Entrepreneur in Residence when a market seems lucrative but it's unclear how to build and launch products and services for such opportunities. Shahid has years of leadership experience as a co-founding startup CTO for multiple venture-backed companies, business unit CTO and EiR, and public company CTO helping transform product teams from marginal to high performance. His software/hardware engineering and cybersecurity body of knowledge is up to date because he rolls up his sleeves to create code when appropriate & dive into system architecture and design when required. He also conduct technology due diligence exercises for corporate acquisition or product integration requirements.
RELATED ARTICLES

Most Popular

CATEGORIES