Glossary · 5 · Building with AI: APIs, tools and agents
AI agent
Also known as: Agentic AI, Autonomous agent
An AI agent is a system in which a language model pursues a goal over several steps on its own — planning, calling tools, observing the results and deciding what to do next — until the task is done or it needs human input.
- Intermediate
- Technical writers
- Technical marketers
- Technical project managers
- Developers
In one sentence
AI agents explained: language models that plan, use tools and act over several steps — what they can do and where humans must stay in control.
Example
A documentation agent receives a change request, finds affected topics, drafts updates in a branch, runs the style checker and opens a review request for the writer.
Explained in context
This term is a station of a mini course
- When an AI agent crosses the line: detect, contain and restore operationsStation: The boundary in three partsStart the course
Context cards connect this term with others to answer one question. Also in British English and German.
- Information user: the functional roleWhose questions drive an adaptive semantic architecture — and what may the system do with them?UKDeutsch
- AI agent as proposer: the functional roleWhat is the role of the AI agent in an adaptive semantic architecture — tool, source of evidence or decision maker?UKDeutsch
- From knowledge graphs to adaptive semantics: the pathway mapHow does a DITA and iiRDS knowledge architecture get from governed semantics to a model that learns from AI agent use — and who does what on the way?UKDeutsch
- The machine is a network, not a monsterWhen an AI answer is wrong, why is “the machine got it wrong” no diagnosis?UKDeutsch
- From job groups to rolesWhy does a grid of job groups — writers, marketers, project managers, developers, the reader, the AI agent — show who is involved but not who does what?UKDeutsch
- When an AI agent crosses the line: detect, contain and restore operationsWhat has to be in place when an AI agent does something it isn’t allowed to do — in IT systems, IoT fleets, OT and on machines — and how is it detected, stopped and released again?UKDeutsch
- Permission, capability, mandate: the three parts of an AI agent’s boundaryWhich boundary does an AI agent cross when it “goes too far” — and why does it matter which one?UKDeutsch
- In-band and out-of-band supervision of an AI agentWhat is the difference between in-band and out-of-band supervision of an AI agent?UKDeutsch
- Indirect prompt injection in agent chainsHow does indirect prompt injection travel through a chain of AI agents, and what changes when the chain gets control points?UKDeutsch
- AI systems and the machine: scope and out of scope of every levelWhere do a company's AI systems end, where does the machine reach into them, and what lies outside every configuration?UKDeutsch
- Workflow-scoped AI and open-ended agentsWhat is the difference between AI embedded in a defined workflow and an open-ended AI agent?UKDeutsch
- The agent master record as a documentation objectWhat should an agent master record for an AI agent contain?UKDeutsch
- Agent workloads vs. microservices and batch jobsHow does an AI agent workload differ from a microservice or a batch job in operations?UKDeutsch
- Control points for AI agentsWho controls what an AI agent is allowed to do?UKDeutsch
Why it matters on your learning path
- Technical writers: Agents will take over routine content maintenance; writers define the rules, review the output and own the result.
- Technical marketers: “Agentic” is a hyped label. Ask what the system decides on its own and what it can change.
- Technical project managers: Define scope, permissions, stop conditions and approval points before an agent touches production systems.
- Developers: Start with narrow, well-evaluated agents; log every step for traceability.
Agent vs. assistant
An assistant answers one request at a time and the user decides the next step. An agent chains steps itself. Many products sit in between — see agentic workflow.