VA seeks information on new AI interface and API for workforce

The Veterans Affairs Department is requesting information about front and back-end-based artificial intelligence for its next phase of AI deployment, according to contracting documents posted Friday and updated Monday.

Specifically seeking a new user interface and API services to large language models, the document said the potential acquisition would expand AI access across the VA’s workforce and support assistive, collaborative and agentic AI capabilities.

“VA is committed to harnessing AI to improve Veterans’ lives, delivering faster, higher-quality, and more cost-efficient services, while upholding the highest standards of safety, security, and equity,” the RFI said. 

The department wants to expand its current AI offerings with a direct contractual relationship for commercially available tools with usage-based pricing while complying with security requirements for sensitive information.

“By deploying commercially available AI platforms from U.S.-based Frontier AI Labs at scale across one of the federal government’s largest workforces, VA contributes to the national objective of demonstrating that American AI leadership translates into measurable improvements in public services,” the RFI said. 

The document detailed how three employees could potentially find value in assistive, collaborative and agentic AI uses: a knowledge worker, like an oncology clinician, could use the tool to look up drug interactions, summarize patient history or generate take-home materials. An operator, like a housing services manager, could summarize regulations, map local housing resource availability, and process referrals. Or, an engineer or quantitative analyst could use it for coding help or build test cases.

But the VA “recognizes that software alone is insufficient to apply AI to high-impact internal and external services, at scale, safely and responsibly,” it said.

The department is also looking for AI literacy and safety products to accompany the AI, as well as potentially placing AI engineers and data scientists embedded within VA operations and environments.

“VA is building enterprise-grade AI infrastructure to support the full spectrum of AI interaction, from assistive to collaborative to agentic, spanning compute, foundation models, data integration, orchestration, and user-facing surfaces, with governance embedded at every layer,” the document said.

In addition, the RFI contained an update on the department’s AI usage.

VA GPT, the in-house generative AI chat tool co-developed with Microsoft, has had 95,000 sign-ups to date since its launch last year and has over 5,500 daily users as of Wednesday, it said. 

On the healthcare side, there was a measurable 22% decrease in mortality associated with the implementation of an AI-assisted clinical decision support tool, and a 21% increase in adenoma detection with AI-powered colonoscopy devices, the document said.

The VA also currently uses Microsoft Copilot and has OneGov licenses for Anthropic’s Claude for Government, Google’s Gemini for Government and OpenAI’s ChatGPT Enterprise FedRAMP. It also has LLM API services through VA Enterprise Cloud, including AWS Government, Azure Government and Azure Commercial. 

The VA said there is more to come and welcomes “constructive feedback” on its plans.

“VA is actively piloting tools from multiple Frontier AI Labs to evaluate user experience, performance, and value at enterprise scale,” the document said. 

There is no incumbent contractor for this effort, and the deadline to respond to the RFI is June 9.