The San Francisco-based Salesforce said that the Defense Department and the Army’s Human Resources Command approved its autonomous AI agents for a range of mission sets involving sensitive military information.

In Wednesday’s announcements, the company explained that its Missionforce National Security platform received Impact Level 5, or IL5, authorization, meaning it is “trusted” to address the “most sensitive, unclassified missions.”

Kendall Collins, the chief executive officer of Missionforce and Government Cloud at Salesforce, said the platform is “modernizing government IT” and that they’re “building the trusted AI backbone for America’s national security.”

According to the announcement, Missionforce is the DoD-authorized variant of the commercial Agentforce platform, which is a network of AI agents designed to support analysis and decision-making.

This is the latest application of the technology since Salesforce won the $5.6 billion contract to build Missionforce in January. In early July, Salesforce said the Air Force deployed Missionforce with the 441st Vehicle Support Chain Operations Squadron to help manage its $13.5 billion fleet of over 84,000 vehicles across nearly 389 locations.

The company called the approval a “milestone” because it moves the national security and defense entities “beyond predictive AI and chatbots” and onto “autonomous AI execution.” The AI agents will have access to controlled unclassified information as well as unclassified National Security System data.

While Missionforce is tailored for government use, the underlying technology is based on Salesforce’s commercial Agentforce 360 platform, which the company introduced in late 2025.

Ryan Von, a writer for Salesforce Ben, a publication that covers the company’s products, described Agentforce as Salesforce’s “boldest leap into AI,” and called the latest 360 variant, “the most powerful version of the tool yet.” He said it’s marketed as able to deliver “accurate, context-rich responses across text and voice.”

In an early review of the commercial platform, Von said demonstrations were promising. Still, he suggested the AI performed best on routine, keyword-based requests while struggling with more complex questions, such as personalized recommendations.

For the Army HRC, the company said AI agents will “provide around-the-clock support, respond to routine inquiries, summarize case histories, and surface relevant policy and career information from approved Army sources.”

According to the company, HRC handles more than 1,500 personnel cases each day. Salesforce estimates the AI agents could save roughly $6 million annually while handling more than 55 million conversations a month.

Bill Pessin, SVP of Missionforce National Security at Salesforce, said that the platform will allow the human staff to “focus more on the complex, career-defining decisions that directly affect the readiness and lives of our Soldiers and less on the mundane tasks that can be handled by AI.”