What is AI agent?

An AI agent is a language model placed in a loop with tools and given autonomy to work toward a goal. Instead of answering a prompt in one shot, an agent plans, takes an action by calling a tool, observes the result, and decides what to do next, repeating until the task is done or it needs help. That action-observation loop is the defining feature: it lets the system search the web, query a database, edit files, or call an API and then react to what it finds. Agents are only as capable as the tools and context available to them, which is where the Model Context Protocol fits, MCP gives an agent a standard way to connect to external tools and data without bespoke integration code. Two other ingredients matter for real-world agents: good context engineering, so each step sees the relevant information, and memory, so the agent does not relearn the same facts every session. A coding agent is one common specialization; broader agents handle research, operations, customer support, and multi-step workflows that mix many tools.