# Semantic Layer

> A semantic layer defines business metrics, dimensions, and joins once, so every query and every AI agent computes the same number the same way.

A semantic layer is the place where a metric is defined once: what revenue means, which tables it comes from, how the joins work, and who may see it. In 2026 it became the main accuracy lever for AI analytics, because an agent grounded in governed definitions produces an agreed answer rather than a plausible one.

**A semantic layer is the layer between your raw tables and the people asking questions, where business metrics, dimensions, joins, and access rules are defined once and reused by everything downstream.** It is the difference between a number and an agreed number.

Its role changed in 2026. For years a semantic layer was a convenience for dashboards. Now it is the main accuracy control for AI analytics, because it is the only thing standing between a language model and a metric it will otherwise define on the fly. [ThoughtSpot Spotter](/tools/thoughtspot-spotter) builds its whole approach on this: instead of generating free-form SQL, it translates a question into search tokens grounded in its governed semantic layer and compiles those into SQL that enforces join logic, hierarchies, and security. [Databricks Genie](/tools/databricks-genie) reaches the same place through curated catalog context, and Databricks' own guidance is that SQL expressions and verified example queries beat plain-text instructions, with metric views used to simplify the model.

The practical version for a small team is less formal but the same idea: a documented set of core tables, column descriptions that say what a field means, and a folder of verified queries that define your key metrics. That collection is a semantic layer in everything but name, and it is what makes [text-to-SQL](/glossary/text-to-sql) dependable.

The tool-by-tool view is in [the best text-to-SQL tools in 2026](/guides/comparisons/best-text-to-sql-tools-2026), and the workflow around it is [Claude for data analysis](/guides/analytics/claude-for-data-analysis).

---

_Source: https://agentscamp.com/glossary/semantic-layer — Term on AgentsCamp._