# Agentic AI

> Agentic AI refers to systems that plan multi-step actions, call external tools, and operate with some autonomy toward a goal, rather than producing a single response to a single prompt.

Category: Architecture
Source: https://impetora.com/glossary/agentic-ai
Part of: Impetora AI consulting glossary (https://impetora.com/glossary)

## What is Agentic AI?

An agentic system pairs a language model with a tool registry, a planning loop, memory, and an exit condition. The model decides which tool to call, observes the result, and continues until the goal is reached or a budget is hit. Patterns include ReAct, plan-and-execute, multi-agent orchestration, and supervisor-worker. Agentic systems are powerful but harder to evaluate, debug, and constrain than single-shot calls. Production deployments wrap them with guardrails, budget caps, action allow-lists, and human approval gates.

## How does Agentic AI apply to enterprise AI?

Enterprises deploy agentic AI for ticket resolution loops, claims handling, sales follow-up, internal research, and any workflow that requires multiple tool calls. The EU AI Act treats some agentic behaviours as higher risk because actions affect external systems.

## Related terms

- [Tool Use](https://impetora.com/glossary/tool-use) - Tool use is the capability of a language model to invoke external functions, APIs, or services as part of producing a response.
- [Function Calling](https://impetora.com/glossary/function-calling) - Function calling is the specific implementation of tool use where the language model emits a structured JSON object matching a function signature, which the host application then executes.
- [Guardrails](https://impetora.com/glossary/guardrails) - Guardrails are runtime checks placed around an AI system to constrain inputs, outputs, and tool calls within safety, compliance, and business policy.
- [Multi-modal AI](https://impetora.com/glossary/multi-modal-ai) - Multi-modal AI refers to systems that can process and generate more than one type of input or output, such as text, images, audio, and video, within a single model or pipeline.

## External references

- [Yao et al., ReAct](https://arxiv.org/abs/2210.03629)

---

Impetora is a custom AI consultancy and solutions partner for enterprises in regulated industries. Submit a project at https://impetora.com/intake.