Agentic system
An AI system that plans multiple steps and calls tools or other systems on its own (searching a database, calling an internal API, invoking another AI system) rather than only producing a single reply. A successful injection against an agentic system can trigger a chain of tool calls instead of one bad sentence, which is why excessive-agency findings on agentic systems must trace what the system did, not only what it said.
Defined in 2 GAGE programs, which carry 2 distinct definitions of it. The wording above is taught in AI Governance: Applied Mastery.
How each discipline defines it
The same term does different work depending on who is using it. These are the definitions as each program teaches them, unedited.
An AI system that plans multiple steps and calls tools or other systems on its own (searching a database, calling an internal API, invoking another AI system) rather than only producing a single reply. A successful injection against an agentic system can trigger a chain of tool calls instead of one bad sentence, which is why excessive-agency findings on agentic systems must trace what the system did, not only what it said.
Anthropic's umbrella term for the whole family of designs along the range from prescriptive, predefined workflows to fully autonomous agents (Anthropic, 2024).
Where it is taught
The exact lessons this term appears in. The first 7 topics of every program are free with a free account.
- AI Agents vs. AI Assistants: Understanding Autonomy · Agentic AI and Workforce Integration, AI Literacy & Professional Conduct
- Red-teaming your own system: attacks a motivated user will find · Evaluation and Trust, AI Governance: Applied Mastery
Terms it appears with
Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.