Why the ‘agentic enterprise’ is the next evolution of federal AI

The conversation in federal IT around artificial intelligence has matured significantly, moving past simple chatbots and basic task automation. For IT leaders responsible for national security and large-scale operations, the next frontier is the agentic enterprise: a fundamental shift from isolated tools to integrated systems where autonomous agents act as a force multiplier for the human workforce. In a new podcast produced by Scoop News Group, Mia Jordan, public sector strategist at Salesforce and former CIO at the Departments of Agriculture and Education, argues that the government’s current fragmented digital model is more than a nuisance; it is a structural weakness that prevents agencies from meeting their full mission potential.

Much of this efficiency gap stems from what Jordan calls “swivel servants” — personnel forced to manually bridge the gaps between disconnected systems. Jordan noted that government employees often switch between applications more than 1,000 times a day, a repetitive cycle that drains productivity and introduces human error. This friction naturally occurs as work moves from one functional area to another, leading to information getting lost in handoffs or buried in unofficial channels like emails, spreadsheets and manual screenshots. For agencies managing rapid response or complex procurement, this fragmentation is a liability; critical context often sits outside official systems of record, leaving leaders to make high-stakes decisions based on incomplete or delayed data.

Addressing this challenge requires more than just “bolting on” additional AI tools. Jordan suggests that agencies must rethink the structure of work itself, moving toward an environment where AI agents operate within a single, secure platform that coordinates tasks across departments. In this model, AI is not a standalone layer but a core component of a unified work environment. The “gateway” to this shift is a shared operating layer that brings disconnected tools into a common workflow, allowing agencies to automate routine activities — such as summarization, approvals, and task assignments — directly within the daily flow of work.

Despite the clear benefits of this evolution, federal IT leaders remain wary of modernization efforts that require costly “rip-and-replace” overhauls of legacy systems. Jordan points to a more pragmatic path forward through zero-copy data architectures, which allow AI to access insights across multiple systems without the risk or cost of physically moving or duplicating sensitive data. By leveraging zero-copy, agencies can achieve the benefits of an integrated enterprise while maintaining the integrity of their existing data silos. As Jordan puts it, “We’re going to blur the lines a little bit now… because information and data can be shared and become a lot more meaningful to more people.”

Ultimately, the move toward an agentic enterprise approach is about reducing the friction that slows mission execution. By transforming fragmented digital environments into coordinated platforms, agencies can ensure that data, workflows and AI-driven support operate in lockstep. This shift allows leadership to stop worrying about the mechanics of moving work from one system to another and instead focus entirely on achieving mission outcomes. As Jordan concludes, the true promise of the agentic enterprise lies in this newfound clarity: “When work happens in a unified environment, leadership can focus more on the outcomes… because the work is now in a single space.”

This interview is part of FedScoop’s Agentic AI Advantage podcast series, sponsored by Salesforce. Explore more expert perspectives on Agentic AI in government here.

This video podcast was produced by Scoop News Group for FedScoop and underwritten by Salesforce.