Agentic AI and the Rise of the Digital Workforce

Artificial intelligence is now a common topic across the Department of War. Yet when many people picture AI in practice, the mental model is usually narrow: a chatbot or model that answers questions about documents or data.

A user asks a question.
The system produces an answer.
The human continues the work.

That model is useful, but it only addresses a small portion of the work that actually happens across defense organizations.

In reality, most of the effort surrounding analysis has little to do with asking questions. The majority of time is spent preparing data, producing artifacts, and moving work through structured processes such as program reviews, budget analysis, contracting actions, and mission planning.

A new approach to artificial intelligence is beginning to address that reality. Instead of focusing only on models that answer questions, organizations are beginning to deploy agentic systems—software agents that participate directly in the workflows where work occurs.

From AI Assistants to Digital Workers

Most AI systems today operate like assistants. A user asks a question, the system responds, and the user decides what to do next.

Agentic systems introduce a different model.

Instead of a single assistant responding to prompts, work is supported by digital workers assigned to different stages of a process. Each digital worker has a set of services it can perform, allowing it to assist with the tasks that normally occur at that step of the workflow.

In this model:

The agent acts as the digital worker.
The prompts represent the services that worker can perform.
The workflow acts as the orchestration layer that assigns work to the appropriate worker.

This structure mirrors how organizations already operate. Work moves through stages, and different specialists handle different parts of the process. Agentic systems extend that familiar model into software.

Where the Work Really Happens

Across the defense enterprise, analysts spend large portions of their time preparing data before meaningful analysis can even begin.

* Information must be pulled from multiple systems.
* Fields rarely match.
* Formats are inconsistent.
* Spreadsheets require cleaning and reconciliation.

Only after this preparation is complete can analysis begin.

Even then, the work continues. Charts must be generated. Summaries written. Reports assembled. Artifacts prepared for leadership reviews, program oversight, or operational decisions.

These activities are essential because decisions depend on them. But much of the labor involved is mechanical rather than analytical. Historically, large amounts of human effort have been required simply because no alternative existed.

Agentic systems help reduce that burden by allowing digital workers to perform many of these tasks as part of the workflow.

Preparing Data for Analysis

Before analysts can generate insight, raw data must first be transformed into a structured form that can be analyzed. Data engineers often refer to this process as ETL or ELT, but most analysts simply know it as the hours spent cleaning spreadsheets, reconciling fields, and organizing information pulled from multiple systems.

Agentic data transformation allows digital workers to assist directly with this stage of the process.

Users can ingest raw datasets and allow an agent to identify relevant fields, normalize formats, reconcile inconsistencies, and structure the information for analysis. Instead of manually cleaning and restructuring data, analysts can move quickly to the stage where insights begin to emerge.

The human still interprets the results. The digital worker performs the preparation work required to reach that point.

Agentic Reporting

After analysis is complete, organizations must communicate results.

In defense environments this often means producing charts, summaries, and formal artifacts such as briefing materials or reports. These outputs are required for leadership decision-making, oversight processes, and program reviews.

Agentic reporting compresses this stage of the workflow.

Instead of manually assembling reports, users can ask questions about their data and receive generated visualizations, written explanations, and formatted outputs ready for review. The digital worker responsible for reporting can generate the artifacts required to move the process forward.

This allows analysts to focus less on producing slides and more on interpreting the information they contain.

Digital Workers Inside the Workflow

The real power of agentic systems emerges when these capabilities are embedded directly into the workflows where work occurs.

Rather than asking users to leave their work and interact with a chatbot, digital workers are assigned to specific stages of a process. As work moves through the workflow, the appropriate agent becomes active and provides the services relevant to that step.

Each digital worker has a library of services it can perform—drafting content, validating inputs, analyzing data, or assembling artifacts. The workflow itself determines when those services are needed.

The result is a digital workforce aligned with the organization’s operational processes.

Example: A Contracting Workflow

Consider a typical contracting workflow inside the Department of War.

The process begins with defining the requirement. Program teams identify the mission need and begin drafting supporting documentation. The effort then moves through stages such as Statement of Work drafting, compliance and risk review, and final contract package preparation.

Each stage requires documentation, analysis, and supporting artifacts.

1. A digital workforce model assigns specialized workers to these stages.

2. During requirement definition, a digital worker might help analyze prior efforts or summarize mission needs.

3. During SOW or PWS drafting, another worker can help generate initial drafts or ensure required sections are present.

4. At the compliance and risk review stage, a worker can evaluate the document for missing clauses or policy inconsistencies.

5. Finally, a worker responsible for contract packaging can assemble the artifacts required for leadership review and approval.

Human reviewers remain responsible for the decisions. The digital workforce simply performs many of the services required to move the workflow forward.

A New Operating Model

Agentic systems represent an evolution in how organizations use artificial intelligence.

Instead of AI that simply answers questions, organizations are beginning to deploy digital workers that participate directly in operational workflows—preparing data, generating artifacts, and assisting humans as work moves through structured processes.

The goal is not to replace human expertise. It is to remove much of the repetitive effort that surrounds it.

When that happens, analysts spend less time cleaning spreadsheets and assembling reports, and more time focusing on interpretation, judgment, and decision-making.

In complex organizations responsible for large missions and vast amounts of information, that shift can significantly accelerate the path from raw data to informed action.

Artificial intelligence, in this model, becomes something more than a tool.

It becomes part of the workforce that helps organizations do the work required to make decisions.

‍