Log in Request a demo

Glossary

AI Agents.

Also called AI agent, agentic AI, agentic AI systems, autonomous AI agents

In Short

An AI agent is software that uses an AI model to pursue a goal, using data, tools and actions with little step-by-step human input. CISA and its international partners say agentic AI systems can introduce extra cybersecurity risks. Agents are also called agentic AI.

Last updated October 23, 2026
Profile MapGlossary Term
  1. Also Called4other namesAI agent · agentic AI · agentic AI systems
  2. Related Terms3linked entriesGenerative AI · Shadow AI · Non-human identity

What Are AI Agents?

An AI agent is software that uses an AI model to work towards a goal with little step-by-step human input. The UK NCSC describes agentic AI as systems that access data sources, remember context, make decisions, use tools and take actions in pursuit of a goal. They do this without continuous human intervention.

NIST says AI agents can now work autonomously for hours, write and debug code, manage emails and calendars, and shop for goods, among other emerging uses.

How Do AI Agents Work?

The NCSC describes agentic systems as a form of advanced generative AI. NIST notes that how useful an agent is depends on its ability to interact with external systems and internal data.

A simple illustration: an agent given the goal “sort my inbox” might read messages, decide what matters and move or reply to them without being asked each time.

Why Do AI Agents Matter for Security?

An agent can act, not only answer. That changes the risk. CISA and its international partners say agentic AI systems can introduce extra cybersecurity risks, such as an expanded attack surface, privilege creep, behavioral misalignment and obscure event records. Privilege creep means an account slowly collects more access than it needs.

The NCSC adds several concerns:

  • Agents can be permitted to access external systems, data and tools.
  • Behavior can be unpredictable, especially when a goal can be read in a way a human would not expect.
  • Problems can be harder to spot when actions happen faster than people can review them.
  • A particular course of action can be harder to explain.

The NCSC also says agentic systems can inherit known risks of language models, such as jailbreaking and prompt injection. Prompt injection is an input crafted to make a model behave in a way it was not meant to. One documented case is the Hugging Face AI agent intrusion, which Hugging Face says was run by an autonomous AI agent system.

What Should Be in Place Before an Agent Goes Live?

CISA and its partners advise organizations to avoid granting broad or unrestricted access, especially to sensitive data or critical systems. They also advise starting with low-risk and non-sensitive use cases, and accounting for agentic AI in the organization’s security model and risk posture.

The NCSC blog on adopting agentic AI (15 May 2026) turns this into a short order of work. It also says humans stay accountable for the decision to deploy an agent, the access it gets, the safeguards around it and what it does.

Before an Agent Touches Real Systems
  1. Ask if AI is neededCheck whether a simpler, lower-risk process change would do the job instead.
  2. Name the people responsibleSay who owns the agent, approves its access, monitors it, reviews incidents and can stop it.
  3. Limit access and scopeGive the minimum access for the shortest time. Constrain actions and when they can run.
  4. Use temporary credentialsAvoid long-lived credentials. Revoke elevated access when the task ends.
  5. Start with a bounded pilotUse clearly defined tasks and build confidence before widening the scope.
  6. Watch what it doesLook for unusual activity across tools, workflows and connected systems.
  7. Plan for failureCover agent failures, misuse and loss of control in incident response plans.

How Do Agents Connect to Shadow AI and Identity?

The NCSC places agentic systems within generative AI. An agent that staff use without approval is a case of shadow AI. An agent that signs in to systems with its own account or credentials is related to non-human identity. NIST lists AI agent security and identity as a research area for trusted adoption.

Is an AI Agent the Same as a Chatbot?

Not exactly. A chatbot mainly replies to messages. Using the NCSC description, an agent also uses tools and takes actions towards a goal. Some chatbots may be built as agents, but the ability to act is what separates the two.

Is “agentic AI” the Same as an AI Agent?

The terms are often used together. The CISA guide uses “agentic AI systems” for software of this kind, and NIST uses “AI agents”. The CISA guide does not give a formal definition of agentic AI.

Also Called.

  • NameAI agent
  • Nameagentic AI
  • Nameagentic AI systems
  • Nameautonomous AI agents

Questions.

  • What are AI agents?

    An AI agent is software that uses an AI model to work towards a goal with little step-by-step human input. The UK NCSC describes agentic AI as systems that access data, remember context, make decisions, use tools and take actions without continuous human intervention.

  • Why do AI agents matter for security?

    An agent can act, not only answer. CISA and its partners say agentic AI can add risks such as an expanded attack surface, privilege creep, behavioral misalignment and obscure event records.

  • How can organizations reduce AI agent risk?

    CISA and its partners advise avoiding broad or unrestricted access and starting with low-risk use cases. The NCSC advises tightly bounded pilots, named owners, least privilege, temporary credentials, ongoing visibility and incident planning.

  • Is an AI agent the same as a chatbot?

    Not exactly. A chatbot mainly replies to messages. Using the NCSC description, an agent also uses tools and takes actions towards a goal. The ability to act is what separates the two.

Sources

  1. 1 Thinking carefully before adopting agentic AI, UK National Cyber Security Centre (NCSC)
  2. 2 CISA, US and International Partners Release Guide to Secure Adoption of Agentic AI, CISA

In Our Research

  1. 1 Agentic AI and Exposed Secrets: How ARTEX Turns a Leaked Key Into Cloud Access November 2026
  2. 2 Aur0ra Ransomware’s AI Agent Intrusions: A Defender’s Task List November 2026
  3. 3 VAPOR PANDA and ARTEX: How AI Agents Ran a China-Nexus Espionage Campaign October 2026
  4. 4 OpenCTI vs MISP Compared: Run Both Locally and Query Them With an AI Agent October 2026
  5. 5 Turn a Threat Report Into ATT&CK Mappings and Sigma Rules With an AI Agent October 2026
  6. 6 Build a Brand and Phishing Watch With dnstwist, crt.sh and an AI Agent October 2026

Keep Exploring.

Knowledge Hub
101 guide Agentic Threat Intelligence 101 How AI agents plan steps and call tools to do threat intelligence work, where a person stays in charge, and a hands-on series to build your own. Open Article Agentic AI In Cybersecurity: How Autonomous Agents Are Transforming The SOC Agentic AI in cybersecurity explained: how autonomous agents triage alerts, investigate incidents and contain threats in the SOC, and what to govern. Open Article When AI Fights Back: How Attackers Are Using Agentic AI Against Your Organization How attackers use agentic AI across the attack chain, what the first AI-orchestrated espionage campaign showed, and how to defend enterprise AI agents. Open Glossary term Attack Surface An attack surface is everything an attacker could reach to try to get into your organization. Open

Browse The Library.