Glossary / Data & Knowledge

Schema

The agreed shape that says what fields data should have and what belongs in each field.

Updated July 2, 2026

A schema is the shape of information, the blank form that makes structured data possible; the data is the filled-in copy. If the schema says there’s a field called “client_name,” the data might say “Acme Plumbing.” If it says “invoice_total” must be a number, then “$1,250.00” may need cleaning first.

A blank intake form captures the idea. It doesn’t hold the client’s answers yet. It tells you what answers you need: name here, email here, service date here, notes here. If everyone uses the same form, the office processes the work without re-asking what each line means.

Why you care

Ask an AI to summarize some clients and you get a nice paragraph, useful to a person but hard for software to reuse. Ask for a specific schema and you get the same fields every time: client name, issue, priority, owner, next action. That’s why schema matters for JSON, databases, and function calling. For agents it’s also a delegation tool, telling the agent what “done” looks like. Without one you get the “where did they put that?” problem: one person writes “company,” another writes “client,” another writes “account.”