Agent
An AI system set up to take actions toward a goal, not just answer a question. An agent can plan steps, use tools, check the result, and try again, often with little or no input between steps. This is the noun behind agentic AI. The distinction that matters for care and for governance is action: a tool that advises leaves the decision and the act with a person, while an agent does the act, which changes who is accountable when something goes wrong and what oversight the system needs.
A referral office pilots an agent that reads referral letters, finds open clinic slots and books patients into them with no coordinator approving each one. One week it reads an old problem in a letter's history as current and gives a routine patient the last urgent slot, so the patient who needed it waits two more weeks. Because the agent made the booking itself, nobody saw the decision until a clinician asked why. Before go-live, list every system it can write to, decide which actions stop for a person to confirm, and check that the log records why each booking was made.
Terms like this come up in real clinical scenarios across the HelloAI courses: bite-sized modules with verifiable certificates. An account takes one minute, no password needed.
Sign in →