--- title: "What is Unity Catalog Metrics?" url: "https://atlan.com/know/ai-agent/databricks/unity-catalog-metrics/" description: "Unity Catalog Metrics, announced at Data + AI Summit 2026, lets teams define governed KPIs once and query them from SQL, BI, APIs, and agents." author: "Emily Winks" author_role: "Data Governance Expert" published: "2026-06-19" updated: "2026-06-19T00:00:00.000Z" --- --- Defining what a metric actually means is one of the oldest unsolved problems in enterprise data. Three teams ask for "revenue" and get three different numbers, because the definition lives in dashboards, notebooks, and [tribal knowledge](https://atlan.com/know/data-for-ai/tribal-knowledge/) rather than in one governed place. At Databricks Data + AI Summit 2026, Unity Catalog Metrics took aim at this directly: define a KPI once, govern it, and serve the same answer to SQL, BI, APIs, and agents. It is a meaningful step toward consistent business meaning, and it raises a larger question that every enterprise running more than one platform eventually has to answer: where does governed context live when your metrics span Databricks, Snowflake, dbt, and a dozen BI tools at once. --- ## Quick facts | Attribute | Detail | |---|---| | What it is | A governance layer for business KPIs in Unity Catalog: define metrics once as reusable objects | | Announced / status | Databricks Data + AI Summit 2026; capabilities in public preview and beta | | Category | Governed metrics layer / semantic modeling inside Databricks | | Who it is for | Data leaders, data architects, analytics engineers standardizing KPIs across the platform | | Key benefit | One governed definition queried consistently from SQL, BI, APIs, and agents | | Works with | SQL, BI tools, APIs, Databricks agents; feeds the Genie Ontology; Power BI and Tableau import (beta) | | How Atlan complements it | Atlan unifies metric and business definitions across the whole estate, adds lineage, policy, and decision traces, and serves context back to Genie | --- ## Why a governed metrics layer matters for agents The reason Unity Catalog Metrics arrived now is that agents made the cost of inconsistent definitions impossible to ignore. A dashboard with a slightly wrong revenue definition is a known footgun a human analyst can catch. An agent that queries the same ambiguous definition at runtime, then acts on it, scales that error silently. Grounding agents in shared business meaning is the difference between a demo and production, which is why governed metrics now sit at the center of the Databricks agentic strategy. According to the [Databricks Unity Catalog blog (June 2026)](https://www.databricks.com/blog/whats-new-unity-catalog-data-ai-summit-2026), Metrics in Unity Catalog let teams "define your business KPIs like revenue, churn, active users, margin, once as governed, reusable objects, then query them consistently from SQL, BI tools, APIs, and agents." It ships alongside two companion capabilities: a Business Glossary for authoritative definitions of business concepts, and Domains for organizing data and AI assets into business-aligned categories. Together they form a governed [semantic layer](https://atlan.com/know/semantic-layer/) foundation inside the lakehouse. Is your metric layer AI-agent ready? Governed KPIs are the start. Check whether your full data estate can serve consistent, trusted context to AI agents at runtime. Assess Your Readiness --- ## What Unity Catalog Metrics actually does Unity Catalog Metrics turns a KPI into a first-class, governed object rather than a formula buried in a query. Because the definition lives in Unity Catalog, it inherits the same access controls and lineage that apply to tables, and any surface that reads from the catalog reads the same definition. ### Capabilities announced at Data + AI Summit 2026 The 2026 release expanded what teams can model and how those models are authored and served. | Capability | What it does | |---|---| | Define once, query everywhere | KPIs like revenue, churn, active users, and margin become governed objects queryable from SQL, BI tools, APIs, and agents | | Multi-fact relationships | Model metrics across multiple fact tables, beyond single-table measures (public preview in Dashboards) | | Level-of-detail (LOD) calculations | Compute at the granularity you choose rather than the grain of the query | | Parameterized metrics | Metrics that adapt to runtime inputs | | Window measures | Improved support for period-over-period analysis | | Materialization | Precompute results so dashboards and agent queries return faster, without changing how metrics are defined (public preview) | | Authoring | Build models visually in the UI, or let agents draft and suggest definitions | | Openness and import | Open source in Apache Spark and Unity Catalog OSS, Open Semantic Interchange ready, with beta import from Power BI and Tableau | These details are drawn from the [Databricks Unity Catalog blog (June 2026)](https://www.databricks.com/blog/whats-new-unity-catalog-data-ai-summit-2026) and the [Qubika DAIS 2026 recap (June 2026)](https://qubika.com/blog/everything-databricks-announced-dais-data-ai-summit-2026/). As [Tredence (2026)](https://www.tredence.com/blog/beyond-dashboards-how-unity-catalog-metrics-powers-trusted-kpis-everywhere) framed it, the goal is "trusted KPIs everywhere," moving metric logic out of individual dashboards and into a governed, reusable layer. ### How it feeds the Genie Ontology Unity Catalog Metrics does not stand alone. The user-defined semantic foundation it creates feeds the [Genie Ontology](https://atlan.com/know/ai-agent/databricks/genie-ontology/), described by Databricks as "a continuously learned [enterprise context layer](https://atlan.com/know/why-ai-agents-need-an-enterprise-context-layer/) in the Databricks Platform." Where the Genie [Ontology](https://atlan.com/know/what-is-ontology-in-ai/) automatically extracts business meaning from tables, dashboards, queries, and connected apps, Unity Catalog Metrics supplies the human-authored, governed definitions the ontology incorporates. As [CIO.com (June 2026)](https://www.cio.com/article/4186154/from-rag-to-ontology-databricks-bets-on-context-as-the-key-to-trusted-ai-agents-2.html) put it, "one definition feeding every agent means you stop getting three different answers to the same question." The same article notes the prerequisite that should anchor any rollout: ontologies do not guarantee correctness, and weak foundational data "just speeds up your existing mess." --- ## Metrics layer, semantic layer, context layer: how they relate It helps to be precise about three terms that get used interchangeably. A metrics layer governs KPI definitions. A semantic layer is the broader practice of shared business definitions for analytics, primarily for BI. A context layer is broader still: it includes semantic and metric definitions plus lineage, policy rules, [decision traces](https://atlan.com/know/what-are-decision-traces-for-ai-agents/), and runtime delivery to [AI agents](https://atlan.com/know/ai-agent/what-is-an-ai-agent/). The context layer is a superset; it consumes and enriches semantic and metric layer definitions rather than replacing them. For the deeper comparison, see [context layer vs semantic layer](https://atlan.com/know/context-layer-vs-semantic-layer/) and [ontology vs semantic layer](https://atlan.com/know/ontology-vs-semantic-layer/). Unity Catalog Metrics is an excellent governed metrics layer inside Databricks. The question for an enterprise is what happens at the boundary, when "customer churn" depends on a CRM definition in Salesforce, a transformation in dbt, and a metric in a Power BI dashboard, none of which live in Unity Catalog. Inside Atlan AI Labs and the 5x accuracy factor Atlan AI Labs measured a 5x accuracy improvement in agents grounded in a governed context layer. See the experiments and the repeatable playbook. Download E-Book --- ## How Atlan's context layer extends Unity Catalog Metrics Atlan is the context layer for AI: the governed infrastructure that delivers enterprise knowledge to every model, every agent, and every team from a single source of truth. Many enterprises run Atlan alongside Databricks Unity Catalog, Snowflake Horizon, or Microsoft Purview, pulling context from all of them into a unified context layer rather than rebuilding from scratch. The frame is better together: Databricks brings the data and the horsepower, and Atlan brings the meaning, governed and live, across the entire data and AI ecosystem. Genie Ontology and Unity Catalog Metrics are excellent inside Databricks. Enterprises also run Snowflake, BigQuery, dbt, Tableau, Power BI, Salesforce, and SAP. Atlan unifies metric and business definitions across all of them and serves that context back to Databricks Genie. The relationship is additive: Unity Catalog Metrics governs KPIs inside the lakehouse, and Atlan extends governed context across every connected system, then feeds it to Genie. ### Metrics layer vs context layer: complementary roles | Dimension | Unity Catalog Metrics | Atlan context layer | |---|---|---| | Primary role | Governs KPI definitions inside Databricks | Unifies metric, semantic, and business context across the whole estate | | Scope | Databricks platform; imports from Power BI and Tableau (beta) | 80+ connectors across warehouses, BI tools, pipelines, and SaaS | | Lineage | Inherits Unity Catalog governance and lineage | Column-level lineage reverse-engineered from SQL across systems | | Policy and trust | Unity Catalog access controls on metric objects | Policy context, decision traces, and CI-validated context before production | | Delivery to agents | SQL, BI, APIs, and agents inside Databricks; feeds Genie Ontology | Enterprise Data Graph served via MCP server, SQL interface, and open APIs to any agent, including Genie | | Best when | Standardizing KPIs within the Databricks platform | Definitions and context span Databricks plus other systems | ### The four products that extend metrics across your stack **Enterprise Data Graph:** 80+ connectors and column-level lineage build a living graph of assets and relationships across the whole estate. Unity Catalog Metrics governs KPIs inside Databricks; the [Enterprise Data Graph](https://atlan.com/know/enterprise-data-graph/) carries that meaning across every connected system and serves it back to Genie. **Context Agents:** AI teammates that auto-generate descriptions, link terms, infer metrics, and propose ontologies. Per Atlan AI Labs (April 2026), [Context Agents](https://atlan.com/know/context-agents/) have generated 690K+ descriptions, with 87% rated on par or better than human writing, across 50+ enterprise customers. These produce the certified definitions a metrics layer needs in canonical form. **Context Engineering Studio:** Bootstrap, test, and ship context as code, with CI-integrated [evals](https://atlan.com/know/ai-agent-evaluation-benchmarks-and-metrics/) before production. The governed definitions agents depend on are validated here before runtime. For the discipline behind it, see [what is context engineering](https://atlan.com/know/what-is-context-engineering/). **Context Lakehouse:** An Iceberg-native, open-format context store that activates via MCP, SQL, and open APIs. Built on open formats, context stored in Atlan stays portable, not locked to any vendor's schema. This is what makes [the context layer for AI agents](https://atlan.com/know/context-layer-for-ai-agents/) work across vendors. According to [Atlan AI Labs research](https://atlan.com/resources/atlan-ai-labs-ebook/), 83% of AI pilots never reach production, and the gap is context, not the model. A governed metrics layer closes part of that gap inside one platform. A context layer closes it across the estate. For Databricks-specific architecture, see the [context layer for Databricks](https://atlan.com/know/context-layer-for-databricks/) guide. See the Context Layer live Watch how teams unify metric and business definitions across Databricks and the rest of the stack, then serve them to agents. Watch Context Layer Live --- ## Governed metrics are the floor, cross-estate context is the ceiling Unity Catalog Metrics is a genuine advance. Defining KPIs once as governed objects, modeling them richly with LOD calculations and parameterized metrics, and serving them consistently to SQL, BI, APIs, and agents removes a long-standing source of inconsistency inside Databricks. Feeding those definitions into the Genie Ontology gives Databricks agents a consistent foundation to reason on. The ceiling on that accuracy is set by how much of your real business meaning the metrics layer can see. When churn depends on Salesforce, dbt, and a BI dashboard at once, a single-platform metrics layer cannot supply the full definition on its own. That is where a context layer is additive: Atlan unifies metric and business definitions across the entire estate, adds lineage, policy context, and decision traces, keeps them portable on open formats, and serves them back to Databricks Genie. Unity Catalog Metrics governs the KPI inside the lakehouse. The context layer makes that KPI true everywhere the business uses it. Book a Demo --- ## FAQs about Unity Catalog Metrics 1. **What is Unity Catalog Metrics?** Unity Catalog Metrics, announced at Databricks Data + AI Summit 2026, lets teams define business KPIs such as revenue, churn, active users, and margin once as governed, reusable objects in Unity Catalog. Those metrics can then be queried consistently from SQL, BI tools, APIs, and agents, and they feed the Genie Ontology. (Source: Databricks Unity Catalog blog, June 2026) 2. **How is Unity Catalog Metrics different from a semantic layer?** Unity Catalog Metrics is a governed metrics layer: it defines what KPIs mean and serves them consistently across SQL, BI, APIs, and agents inside Databricks. A [semantic layer](https://atlan.com/know/semantic-layer/) is the broader practice of shared business definitions for analytics. A context layer is broader still: it includes semantic and metric definitions plus lineage, policy rules, decision traces, and runtime delivery to AI agents. 3. **What semantic modeling capabilities does Unity Catalog Metrics add?** Databricks expanded metric modeling with multi-fact relationships, level-of-detail (LOD) calculations that compute at the granularity you choose, parameterized metrics that adapt to runtime inputs, and improved window measures for period-over-period analysis. Materialization precomputes results so dashboards and agent queries return faster. (Source: Databricks Unity Catalog blog, June 2026) 4. **How does Unity Catalog Metrics relate to the Genie Ontology?** Metric definitions in Unity Catalog feed the [Genie Ontology](https://atlan.com/know/ai-agent/databricks/genie-ontology/), the continuously learned enterprise context layer in the Databricks Platform. The metrics provide human-defined, governed semantics that the ontology incorporates so Genie agents reason on consistent business meaning. 5. **Does Unity Catalog Metrics work with Power BI and Tableau?** Databricks announced beta support for importing existing models from Power BI and Tableau into Databricks metrics. The metrics implementation is also open source in Apache Spark and Unity Catalog OSS and is built to be Open Semantic Interchange ready. (Source: Databricks Unity Catalog blog, June 2026) 6. **How does Atlan complement Unity Catalog Metrics?** Atlan is the context layer for AI. It unifies metric and business definitions across the entire estate, not just Databricks, adds column-level lineage, policy context, and decision traces, and serves that context back to Databricks Genie and to any other agent through an MCP server, SQL interface, and open APIs. The two are additive: Unity Catalog Metrics governs metrics inside Databricks, and Atlan extends governed context across every connected system. 7. **Is Unity Catalog Metrics generally available?** At Data + AI Summit 2026 Databricks announced several Unity Catalog Metrics capabilities at varying maturity. Multi-fact relationships and materialization were announced in public preview, with other modeling enhancements rolling out. Check the Databricks Unity Catalog blog for the current status of each capability. (Source: Databricks Unity Catalog blog, June 2026) --- ## Sources 1. [What's new with Unity Catalog at Data + AI Summit 2026, Databricks Blog](https://www.databricks.com/blog/whats-new-unity-catalog-data-ai-summit-2026) 2. [Databricks Launches Genie One: All-New Agentic Coworker for Every Team, Databricks](https://www.databricks.com/company/newsroom/press-releases/databricks-launches-genie-one-all-new-agentic-coworker-every-team) 3. [Everything Databricks Announced at the DAIS Data + AI Summit 2026, Qubika](https://qubika.com/blog/everything-databricks-announced-dais-data-ai-summit-2026/) 4. [Beyond Dashboards: How Unity Catalog Metrics Powers Trusted KPIs Everywhere, Tredence](https://www.tredence.com/blog/beyond-dashboards-how-unity-catalog-metrics-powers-trusted-kpis-everywhere) 5. [From RAG to ontology: Databricks bets on context as the key to trusted AI agents, CIO.com](https://www.cio.com/article/4186154/from-rag-to-ontology-databricks-bets-on-context-as-the-key-to-trusted-ai-agents-2.html) 6. [Key takeaways from day two of Databricks Data + AI Summit, SiliconANGLE](https://siliconangle.com/2026/06/17/key-takeaways-day-two-databricks-data-ai-summit/) 7. [Databricks bets on owning the agentic data stack at Data + AI Summit 2026, Moor Insights & Strategy](https://moorinsightsstrategy.com/field-notes/databricks-bets-on-owning-the-agentic-data-stack-at-data-ai-summit-2026/) 8. [AI governance at Data + AI Summit 2026: What's new with Unity AI Gateway, Databricks Blog](https://www.databricks.com/blog/ai-governance-data-ai-summit-2026-whats-new-unity-ai-gateway) 9. [Unlocking Business Context in Data: How Databricks Metric Views Bridge the AI Gap, Aimpoint Digital](https://www.aimpointdigital.com/blog/databricks-unity-catalog-metric-views-cementing-your-business-semantics-to-power-analytics-ai)