Glossary
agentic
An informal descriptor for AI systems that pursue multi-step goals via tool use, planning, and self-correction, rather than single-turn question-answering.
A loose label that distinguishes systems that loop (plan, act, observe, revise) from systems that answer one question and stop. The constituent capabilities are function callingagentsA pattern where a model emits a structured call (function name plus arguments), the runtime executes it, and the result returns as input on the model's next turn. Open full entry , persistent memory, the ability to read the result of an action and decide what to do next, and sometimes the ability to spawn subagents.
The label is fuzzy on purpose. Vendor marketing uses “agentic” for any chatbot that calls a tool; researchers reserve it for systems with genuine autonomous loops; the meaning slides between them in everyday discussion. Anthropic’s 2024 “Building Effective Agents” essay distinguishes simple workflows (deterministic chains) from agents (LLM-driven control flow) and is a useful reference point.
The architectural shift agentic systems require: tool definitions matter as much as system prompts; observability needs traces across many model calls; cost and latencycomputeThe time from request submission to response completion, broken down for LLMs into time-to-first-token and time-per-output-token, the user-facing speed metric. Open full entry budgets per task replace per-call budgets; failure modes shift from wrong answers to unbounded loops and tool misuse.