Autonomy terms

A software system powered by an LLM that perceives its environment, plans actions, and executes them — usually across multiple steps and tools.

AI systems that act with autonomy — perceiving their environment, planning multi-step actions, calling tools, and iterating toward a goal — as opposed to single-turn generative AI that only responds to prompts.

A workflow where an AI agent plans, executes, and adapts a multi-step process — autonomously calling tools, reading data, and looping until the goal is met.

The transition pattern where an AI agent transfers control of a task to a human or to another agent — preserving context, state, and prior actions so the receiver can continue seamlessly.

A workflow where an engineer codes alongside an AI assistant that suggests, completes, and reviews code in real time — distinct from autonomous coding agents that ship PRs without human intervention.

An agent that plans, executes, and finishes a multi-step task without asking for human approval between steps.

An AI tool that suggests changes inline and waits for the user to accept — the human stays in the driver's seat.

A multi-agent architecture where a "manager" or "planner" agent delegates sub-tasks to specialist worker agents — the most common multi-agent pattern in 2026 production systems.

A workflow pattern where the agent pauses for human approval at one or more checkpoints before continuing.

An agent that plans and executes most steps unsupervised but pauses for approval before anything irreversible.

In hierarchical multi-agent systems, the top-level agent that delegates work to specialist sub-agents, monitors progress, handles failures, and aggregates results.

A multi-agent pattern where many similar agents collaborate without a central supervisor — inspired by ant colonies and bee swarms, used for parallel exploration and consensus.