Build Basics / Standard term
Request and response
A communication pattern in which a client sends a request and a server may return a response describing the result.
A request and response is the back-and-forth where one piece of software sends a request asking a server for something, and the server sends a response with the answer. The request names what you want and carries any details the server needs; the response carries the result plus a status code that says whether it worked. Picture a password reset: your browser sends a request that says "reset the password for this account," the server checks the account and emails a reset link, then sends back a response confirming the request was accepted. Every click that loads new content is one of these round trips, usually finished in a fraction of a second.
Builder example
When you wire an AI assistant to a tool, you are asking it to send a request and read the response. If you tell the agent to fetch open to-do items and the response comes back with an error status instead of data, the agent needs to see that status to retry or report the problem rather than inventing items. Designing each call so the response is checked, not assumed successful, is what keeps an automated flow honest.
Common confusion: A response may report an application error, and no response may arrive after a timeout, disconnect, crash, or dropped packet. Some requests can also succeed even when the client never receives the acknowledgement, which makes retries unsafe without idempotency.