The Pentagon has never let an AI agent touch sensitive data. That changed on August 5. Missionforce, Salesforce’s national security unit, got Impact Level 5 authorization for its Agentforce 360 platform — autonomous AI agents can now handle Controlled Unclassified Information and unclassified National Security Systems data. First agentic platform ever cleared at this level.
What IL5 actually unlocks
IL5 is the DoD’s most sensitive tier before classified. Until now, agents were locked out — humans did the paperwork. These agents come in two flavors: assistive copilots that ride alongside staff, and fully autonomous agents that act on their own. Target workloads: defense logistics, recruit onboarding, and the administrative grind that buries military families.
9.2 million people, 1,500 cases a day
Army Human Resources Command is the first deployer, putting 24/7 AI agents in front of 9.2 million soldiers, veterans, and family members. Expected throughput: 1,500-plus automated case summaries daily.
The bigger signal: this sits on a $5.6 billion, ten-year contract awarded in January. Government AI agents just went from pilot curiosity to line-item procurement — and Salesforce, not a frontier lab, got there first.
You Might Also Like
- Agent Builder by Thesys When ai Agents Stop Talking and Start Showing
- Google A2ui Agent to User Interface Finally a Standard way for ai Agents to Show you Things
- Agent Action Protocol aap the Missing Layer Above mcp That Actually Makes Agents Production Ready
- Agent Safehouse Finally a Dead Simple way to Stop ai Agents From Roaming Your mac
- Google is Deploying Gemini ai Agents to 3 Million Pentagon Employees Heres What That Means
Top AI Product 
Leave a comment