A schema is the formal structure of a database: its tables, the fields in each, their types, and the links between them. Wikipedia describes it as "the structure of a database described in a formal language", and as "a blueprint of how the database is constructed".
What a schema says
A schema tells the software that a table called TICKET has a field called NET_WT holding a number, a field called CONTRACT_ID that must match a row in the CONTRACT table, and a date that can't be empty. It enforces rules: no ticket without a contract, no text in a number field.
What it doesn't say
It doesn't say what things mean. Is NET_WT in pounds, kilograms or tons? Is it before or after the deduction for bark? What does ADJ_FLG = 2 signify? Why do some tickets have a second quantity field that is sometimes filled in? Those answers live in people's heads, old emails and vendor manuals.
That gap shows up in AI accuracy. BEAVER, a benchmark built from private enterprise data warehouses, spans 812 tables across 19 domains. Leading systems "achieve only 10.8% accuracy" on it. Given annotations that supplied the missing domain knowledge, accuracy "increases to 30.1%". Its authors name domain knowledge and query complexity as the main challenges.
In a timber business
A timber system's schema can be decades old and extended many times, with fields reused for new purposes and codes that have changed meaning. Two systems can store the same fact under different names and units.
What helps
A short, plain-language description of each important field, kept beside the schema: what it holds, in what unit, and how it is used. It helps new staff, analysts and AI tools alike. Hand-written descriptions drift as the system changes, so newer tools also infer meaning from the data and past queries, with a person confirming what they suggest.
What it isn't
A schema isn't a data model, though it implements one. The model describes the business, and the schema describes one database. In some systems, "schema" also means a named group of tables inside a database.
How Quarri works explains the platform as a layer over existing systems, not a migration.
Sources
- Wikipedia, "Database schema": en.wikipedia.org
- Chen and others, "BEAVER: An Enterprise Benchmark for Text-to-SQL", arXiv 2409.02038, revised 13 May 2026: arxiv.org
Quarri is an AI-native data platform for the timber supply chain. It connects buying, production, sales and inventory for forest management, sawmill, wood products and pulp, paper and packaging operators.