Primer · For Policymakers

What policymakers need to know about agentic AI.

A plain-language guide to the technology behind the headlines. What an AI agent actually is, how these systems work, and the questions Congress and regulators will need to answer. No background required.

10 minute read Updated June 2026 Non-partisan · Non-technical Download PDF

01The shift from answering to acting.

For the last few years, the AI most people encountered did one thing: it answered. You typed a question, it produced text. The human stayed in the loop on every step, reading each output and deciding what to do next.

Agentic AI is different. An agent doesn't just answer. It acts. Given a goal, it can plan a series of steps, make decisions along the way, and carry them out with limited human supervision. It can browse the web, run code, fill out forms, move money, and work with other software systems on a person's behalf.

The simplest way to hold the distinction: the old model is a research assistant who hands you a memo. An agent is a junior staffer you can hand the whole task to.

Answering versus acting A traditional model returns an answer for a human to act on. An agent takes a goal, plans steps, uses tools, and carries out the task, reporting back at set checkpoints. TRADITIONAL MODEL You ask a question It returns an answer You do the work human acts on every step AGENT You set a goal It plans, decides, acts searches · compares · executes The task is done reports back at checkpoints
A traditional model answers. An agent takes the goal and does the task, checking back at the points you set.

That shift is why existing AI policy frameworks don't map cleanly onto these systems. Most were written for static models and human-reviewed outputs. When software starts taking actions in the world, a different set of questions opens up around accountability, identity, and oversight. Those are the questions this primer is built to frame.

02The vocabulary, in one place.

A handful of terms come up in nearly every agentic AI conversation. They sit in a rough hierarchy. Understanding how they relate is most of what it takes to follow a technical briefing.

The vocabulary of AI, as a hierarchy Foundation models are the base layer. Frontier models are the cutting-edge subset. Agentic AI is the shift from models that answer to systems that act, built from tool use, orchestration, and multi-agent design. FOUNDATION MODEL · THE BASE LAYER Large, general-purpose models. The LLM you know is the most familiar kind. Generative creates new content Multimodal text, images, audio Frontier cutting-edge · regulated Every frontier model is a foundation model, but most foundation models are not at the frontier. THE SHIFT AGENTIC AI that plans, decides, and acts within set boundaries, using a model as its reasoning engine. THE MACHINERY THAT MAKES IT WORK Tool use reaching outside itself to act Orchestration wiring models and tools into one system Multi-agent specialized agents working together
The terms sit in a hierarchy. Foundation models are the base. Frontier is the regulated cutting edge. Agentic is the shift to systems that act.

Those three building blocks, in a little more detail:

Tool use
An agent's ability to reach outside itself, calling a search engine, a database, a payment system, or another application to get something done.
Orchestration
Wiring multiple models, tools, and agents together into one coordinated system that can handle a complex, multi-step workflow.
Multi-agent
An arrangement where several specialized agents work together, each handling part of a larger task and passing work between them.

One concept runs underneath all of these and deserves its own line, because it's where much of the policy weight lands: identity and authorization. This is how a system knows which agent is acting, whose authority it carries, and what it has been permitted to do. When software acts on your behalf, that permission question is the whole game. The industry is building the standards for this right now, across open-source foundations, formal standards bodies, and the payments ecosystem. For a plain-language map of who is doing what, see AFI's standards landscape reference.

03How an agent actually works.

Picture a concrete task: "Book me a flight to Chicago next Thursday under $400, aisle seat." A traditional model would draft you an email about flights. An agent does the job.

It breaks the goal into steps, searches for flights, compares options against your constraints, selects one, and completes the booking, checking back with you at the points you've told it matter. Two things define what it can and can't do. First, what it's allowed to act on: which systems and accounts it can touch. Second, what it can see versus what it reports back to you. An agent given a narrow scope and a clear reporting line is very different from one with broad access and little visibility.

There's one more property worth understanding, because it surprises people from a traditional software background. Agents are non-deterministic. Run the same task twice and you may get two slightly different paths to the answer. That's not a bug. It's a consequence of how these models reason. But it's exactly why supervision, logging, and clear boundaries matter more here than in conventional software, where the same input always yields the same output.

"Agent" is not one thing. Systems fall along a spectrum of autonomy, from a tool that acts only when prompted to an actor that pursues open-ended goals on its own. Most of what's deployed today sits in the middle, executing bounded goals and reporting back. Where a system falls on this scale is the single most useful question for a policymaker, because it determines how much human oversight is realistic and how high the stakes run.

The autonomy spectrum Agents range from reactive tools that act only when prompted, through assistants that suggest, to systems that execute within set bounds, to autonomous actors that pursue goals with little supervision. As autonomy rises, the human oversight burden falls and the policy stakes climb. LESS AUTONOMOUS MORE AUTONOMOUS 01 · Tool Acts only when you prompt it, one step at a time. 02 · Assistant Plans and suggests, but waits for you to approve each action. 03 · Bounded Executes a goal within limits you set, and reports back at checkpoints. WHERE MOST AGENTS SIT TODAY 04 · Autonomous Pursues open- ended goals with little supervision. Largely still ahead of us. HUMAN OVERSIGHT NEEDED PER ACTION High: you approve each step Low: you set bounds, then trust The right level of oversight depends on where a system sits, and so does the policy question.
Autonomy is a spectrum, not a switch. As systems move right, the human oversight needed per action falls and the policy stakes climb.

04The policy questions that follow.

You don't need to resolve these to be conversant in them. Naming the questions clearly is most of the work. Each is an area where existing frameworks were not built for software that acts.

Accountability

When an agent takes an action that causes harm, who is responsible? Responsibility is distributed across the model developer, the company that built the application, the company that deployed it, and the user who set the goal, and the harm can be hard to trace to any single one. Scholars call this the problem of many hands. Existing liability law wasn't written with an autonomous actor in the middle.

Identity & authorization

How do systems verify that an agent is who it claims to be and is authorized to act for a given person? This is the foundation everything else rests on, and the infrastructure for it is still being built.

Interoperability & access

Will agents be able to work across different platforms and services, and on what terms? Technical standards, security, and business considerations all bear on this. How it resolves shapes how easily new entrants can build and how much choice users have.

Workforce

As agents take on tasks rather than just answer questions, what happens to the work people do? Estimates of the scale vary widely and the picture differs by sector and role, with some work augmented, some automated, and new roles created. It is the question constituents feel most directly, and the one where good data matters most before policy is set.

Security & privacy

An agent with broad access and the ability to act is a powerful tool and a new attack surface. What controls, logging, and data protections should govern systems that can act on sensitive accounts?

Oversight & supervision

What does meaningful human oversight look like for a non-deterministic system operating at machine speed? The field distinguishes a human in the loop (approving each action), on the loop (monitoring and able to intervene), and in command (setting the boundaries up front). The hard question is which model fits which use, and where a human checkpoint is essential versus unworkable.

05Where AFI fits.

The Agentic Futures Initiative is a cross-industry coalition of the companies building and deploying agentic AI. We don't ask policymakers to take our positions on faith. Our role is to make sure the people writing the rules understand the technology well enough to write them right, through briefings, filings, and plain-language resources like this one.

If a briefing, a deeper explainer on any of these questions, or a technical walkthrough would help your office, that's exactly what we're here for.

This primer is a non-partisan educational resource. Definitions reflect consensus industry usage as of June 2026. The regulatory definition of "frontier model" varies by jurisdiction and is tied to compute thresholds in the EU AI Act and several U.S. state laws. For a briefing or to go deeper on any section, contact info@agenticfuturesinitiative.org.

Want this as a briefing?

Request a Briefing View All Resources