DIA Seeks AI and Machine Learning Prototypes to Strengthen Military Intelligence
The US Defense Intelligence Agency (DIA) is calling on industry leaders to submit innovative prototypes that enhance decision-making capabilities for warfighters using artificial intelligence (AI) and machine learning (ML). The agency is looking to close critical technology gaps through its latest solicitation.
On April 15, 2025, DIA issued solicitation HHM402-23-SC-0002-0001 under the NeedipeDIA Commercial Solutions Opening (CSO), titled Call #CSO-0025-001. The initiative falls under the purview of Task Force SABRE, a specialized, limited-duration organization tasked with accelerating the integration of AI across DIA systems.
Focus on Three Key Capability Gaps
The solicitation outlines three primary areas where DIA is seeking innovative solutions:
1. Enterprise Multi-Model Large Language Model (LLM) Applications:
DIA wants to utilize existing and future large language models to generate human-like responses from vast amounts of data. Proposed solutions should feature a user-friendly interface, support for switching between models, document upload functionality, and a prompt output cache.
Additionally, systems must be secure, scalable to over 2,000 concurrent users, and easily extendable to support Retrieval-Augmented Generation (RAG), AI agents, and media generation.
2. Multi-Source Data Retrieval and Pre-Processing
Solutions should retrieve data from geographically dispersed sources, pre-process it for analysis, and manage storage across various security classifications. Proposals must include cybersecurity measures, fine-grained data access controls, cross-domain data transfer capabilities, and support for building vector databases for use with LLMs.
3. AI-Enabled Document Generation and Assessment
DIA also seeks tools that use LLMs to create and assess standardized intelligence documents. These tools should be able to parse documents using domain-specific knowledge, evaluate LLM outputs, offer feedback mechanisms, and feature both automated and manual data retrieval functions.
Also read: BD Launches AI-Powered HemoSphere Alta to Advance Hemodynamic Monitoring
Security and Scalability Are Essential
All proposed solutions must meet strict security criteria. Organizations must be cleared to work with classified data up to and including Top Secret, and their personnel must hold the appropriate security clearances. Access to classified facilities is also required.
The ability to scale these technologies across the agency and adapt to different levels of data sensitivity is a key factor in the selection process.
Proposal Submission Deadlines
Interested companies must submit a two-page white paper outlining their solution no later than 30 April 2025. White papers should be emailed to: Ideas2Action@dodiis.mil.
Selected applicants will be invited to submit full proposals, which will be due by 23 May 2025. The DIA expects to announce contract awards on or around 16 June 2025.
Full details and official documentation are available at: https://sam.gov/opp/066a8698563f45ad9424dd3355bbb623/view
