Autonomous AI refers to systems that pursue goals and take actions without step-by-step human direction. In practice, autonomy is granted in degrees, bounded by permissions, budgets and human approval for consequential actions.
An AI agent framework is a library or platform that provides the building blocks for agents: the execution loop, tool integration, memory, state management, error handling and tracing.
Multi-agent AI is an approach where several AI agents, often with different roles and tools, work together on a task, coordinating through an orchestrator or by passing work between them.
Agent planning is how an AI agent breaks a goal into steps, decides their order and revises them as it learns from results. It is what separates a multi-step agent from a single tool call.
Agent memory is how an AI agent retains information, both within a task and across sessions, ranging from the working context window to persistent stores of facts, past actions and learned preferences.
Function calling is the API feature that lets a language model return a structured request to run a function you have defined, with arguments matching a JSON schema you supply.
Tool calling is the mechanism that lets a language model request actions from external systems, such as searching a database or sending an email, by emitting a structured request that your code executes.
An agentic workflow is a process where AI carries out or chooses multiple steps towards a goal, ranging from fixed chains of model calls to fully agent-driven execution.
Agentic AI describes systems that act with a degree of autonomy towards goals, planning steps and using tools, rather than simply responding to each prompt.
An AI agent is a system that uses a language model to pursue a goal over multiple steps, deciding for itself which tools to use and when the task is complete.