Agents chain LLM calls with tool use — search, code, email, payments — and loop until the goal looks done. That autonomy multiplies both capability and blast radius: an injected instruction can now trigger real side effects, not just a bad reply.
GlossaryPart of the BotDoors AI Safety & Compliance Glossary · 37 terms.
Safety
Agentic AI
Agentic AI describes systems that plan multi-step actions and use tools autonomously to reach a goal.