Putting AI Agents to Work Across Federal Missions

Federal agencies are moving beyond personal AI assistants to introduce coordinated agents into high-volume workflows that directly support cyber response, service delivery, compliance, and mission operations. Doing so responsibly requires agencies to understand where agents can deliver value, how they are making decisions, and when people should intervene.

In a preview to MeriTalk’s Aug. 27 Mission Room forum, “Agentic AI and the New Federal Workflow,” MeriTalk sat down with Dave Erickson, public sector distinguished architect at Elastic, to discuss how agencies can identify promising workflows, set boundaries for agent autonomy, and evaluate AI-assisted decisions. Erickson also explored how search, observability, real-time analytics, and higher-level agents can help agencies reduce risk while keeping human attention focused on mission outcomes.

MeriTalk: What characteristics make a federal workflow a strong candidate for coordinated AI agents?

Erickson: We should differentiate between personal software and official workflow process automation. We have all fallen in love with the little agent that tidies our email, summarizes a rambling status update, or otherwise helps us throughout the day. That’s the personal software approach.