The Difference That Matters

A chatbot receives a question and returns an answer. An agent receives an objective and decides what to do.

The difference may seem subtle. Ask a chatbot to "write Ana’s letter" and it will write with whatever is in front of it, including making up what’s missing, because the only thing it knows how to do is produce plausible text.

An agent opens the field report, notices that school attendance is missing, checks the registry, writes, checks the text against the source, and stops to escalate because the child moved to another city.

The text may even come out looking similar. What changes is the number of decisions it made on its own, and each of those decisions is a point where you can gain autonomy or lose control.

Everything called agentic architecture comes down to managing a four-stage cycle: think, act, observe, repeat. The design questions are always the same four: what information does it think with, what tools does it have to act, what can it observe about the result, and when does the loop stop.

That last one is the most often forgotten. An agent without a written stop condition works like an open tab.