Accelerating Federal IT modernization by streamlining deployment of private cloud AI agents across government networks

Share “A Game Changer in Federal Agentic AI Modernization” on TwitterShare “A Game Changer in Federal Agentic AI Modernization” on FacebookShare “A Game Changer in Federal Agentic AI Modernization” on LinkedIn

I’m finding that technologists in and out of government are excited and curious about agentic AI and the impact it could have on streamlining software development. Agentic AI refers to a type of artificial intelligence designed to not only respond to chatbot queries but also to perform tasks autonomously, taking actions based on user input or environmental conditions. It’s worth taking a moment to recognize the significance of this development, particularly since we’re already comparing it with “traditional” — a mere three or four years old — generative AI. 

Generative AI can operate on a “read-only” model—responding with answers, suggestions, or predictions based only on what it knows from when it was trained (which may have been just months ago). Agentic AI elevates this interaction by being able to check databases and other sources of information and executing tasks autonomously. The impact on productivity and operations across agencies could be game-changing, especially for the government software developers working to modernize federal agency environments and achieve the transformational objectives set by the Administration’s AI Action Plan.

Building an AI agent can be complex but now there are tools on the market that help software developers create AI agents quickly, freeing developers to focus on more complex problems and new innovations.

What are Agent Builders?

Running in a private cloud operating model in order to protect data privacy and sensitive workflows, an Agent Builder offers the potential to transform the way federal agencies build, deploy, and use agentic software for the good of their agencies and the citizens who depend upon them.

Agent Builders bring the ability to chain together separate machine learning-driven operations to perform many of the most critical—and time intensive—tasks involved in application development or systems integration. Once chained together, the software can perform individual, discrete actions, saving users time.

In this way, Agent Builders enable government agency IT teams to quickly focus on the key priorities intrinsic to their agency mandates, assembling increasingly ambitious chains of what otherwise would be time-intensive manual or machine-to-machine activity, so that the Agentic AI itself performs these functions.

The result can be Promethean: The agent interacts across different disparate systems, runs iterative A/B tests, trials multiple versions, performs compatibility testing, and conducts security tests and compliance checks. The software itself becomes quasi-autonomous. Enterprise software applications can now be fashioned with greatly reduced human effort. This drives efficiency and productivity, allowing users to focus on more complex problems and faster innovation cycles.

Use Case Scenarios

It’s hard not to marvel at the seemingly limitless potential scenarios for both developers and end-users of federal services, on the web or within government IT environments. While agentic AI must always operate within appropriate guardrails, especially in military and national-security environments, there are myriad possibly transformative use cases for government agencies. Let’s look at three.

Data Scientists: Private Cloud Agent Builders are ideal for data scientists and the infrastructure administrators who support them. These Agent Builders have baked in model governance, giving data scientists the ability to rapidly experiment and iterate on training, fine tune their models, and share the results with teams that have appropriate data access. Non-specialist teams can then use the models to build Agents in a user-friendly interface, enabling data scientists to focus on their area of special expertise. The private AI Agent Builder can do all this work in protected networks, avoiding public cloud vulnerabilities.

Contact Centers: Agencies can create a “customer service representative-in-a-box” while securing the intellectual property (IP) that protects both the government agency and the privacy rights of citizens engaging with it. This approach offers federal organizations faster ways to elevate customer satisfaction and streamline resolution, making agencies more efficient and productive. Developers can deploy the platform using natural-language descriptions of what agency services must be supported and what type of customized user-service experience is expected.

Military (and Other) Logistics: For any large federal enterprise that needs to handle complex logistics involving a wide variety of stakeholders and role players, Agentic AI will have significant impact on operations. The power of Agentic AI is its ability to connect disparate, previously unintegrated systems – and it can even file tickets, provide a clear plan of action that it will take, and ask for permission from a human supervisor before proceeding. For example, a very heavy, human-centric chain of events is currently required to manage the movement of military hardware, support supplies, and service personnel before, during and after a crisis in remote regions of the world. With guidance and guardrails from stakeholders across organizations, an Agent Builder can create the necessary chain of fully or semi-autonomous response steps in the software itself — along with identifying appropriate signals and triggers to activate them as an emergency arises.

Agent Builders and Private Cloud

Private cloud Agent Builders can operate within an agency’s assured and trusted security perimeter and IT environment, rather than relegating developers to building such AI agents in less secure public cloud environments.

This secure, compliant approach affords agencies the confidence to create any required number of agents, across any number of projects and departments. Users are empowered by having assurance they can create in a trusted environment, without exposing their data, prompts and algorithms to potential exploitation.

Private cloud helps ensure that protection: It provides an IT environment that emphasizes the privacy of the data being used, security for the agency processes and intellectual property (IP), and compliance with federal mandates. All requisite regulatory procedures and controls are implemented around and within the mission-critical innovation driving their organizations forward—including the agency’s powerful new AI agents.

For information on one leading approach, learn more about Broadcom’s VMware Agent Builder