Agents
Agentic AI
AI systems that autonomously plan and execute multi-step tasks using tools, memory, and loops, going beyond a single-turn chat response.
Agentic AI is a paradigm in which AI systems pursue a goal autonomously across multiple steps, rather than returning a single response per prompt. The system plans what to do next, calls tools (search, code execution, APIs), observes the results, and loops until the task is complete.
Unlike a generative chat assistant that produces text on request, an agentic system is proactive: it makes decisions and acts in pursuit of a goal. The core building blocks are a large language model as the reasoning engine, a set of tools (see tool use), memory and context, and a runtime loop that drives execution. A single instance of such a system is called an AI agent.
In 2025–2026 agentic AI became the dominant frontier theme, with production use in coding, customer support, and research. Open challenges remain reliability over long trajectories, error compounding across iterations, and the safety of autonomous action.