# Overview

## What are Governed Metrics?

**Governed Metrics are the official, trusted definitions of business metrics in Athenic AI.** They turn common questions like “What is ARR?” or “How do we calculate retention?” into **one consistent definition** that everyone in the organization can use across dashboards, reports, and AI-driven analysis.

Governed Metrics are built on a semantic modeling layer meaning each metric is defined with:

* A clear calculation (the “math”)
* The underlying data sources (models/tables)
* Consistent filters and business logic
* Supported dimensions, entities, and time grains (where applicable)

In short: Governed Metrics are how you standardize metric definitions so the numbers are reliable and repeatable.

### What issues does it solve?

Without governance, metrics tend to fragment:

* Different teams calculate the “same” metric differently
* Dashboards disagree
* Analysts spend time reconciling definitions instead of answering questions
* AI tools return inconsistent results depending on how a question is phrased or which dataset is queried

Governed Metrics solve this by providing a **single source of truth** for KPI definitions. That means:

* **Consistency:** the same metric produces the same result everywhere
* **Trust:** stakeholders know what a metric includes (and what it doesn’t)
* **Scalability:** new dashboards, ad hoc analysis, and AI queries all reuse the same definitions
* **Observability:** metric logic is explicit and reviewable

### How Governed Metrics enable AI-powered Insights

Governed Metrics provide a structured, trusted layer that allows AI to answer business questions using the same logic your team has approved.

#### **Deep Research (Exploratory & Agentic)**

{% hint style="info" %}
**Deep Research does not currently support Governed Metrics. This is in active development.**
{% endhint %}

With Governed Metrics, Deep Research can:

* Explore trends and drivers using consistent KPI definitions
* Break metrics down by supported dimensions (e.g., region, product, segment)
* Compare time periods using the correct grains and date logic
* Generate deeper investigations without “making up” metric logic on the fly

Because the metrics are governed, the AI can focus on **insight generation** (what changed, why, what’s driving it) instead of creating and inferring its own definitions and KPIs.

#### **AI Analyst Queries**

When someone asks:

* “What was revenue last quarter?”
* “Why did churn increase in December?”
* “Show me retention by cohort”

Athenic AI can map those questions to Governed Metrics and return answers that are:

* Consistent with dashboards and reporting
* Reproducible by analysts
* Grounded in your approved semantic logic

### What to expect next

In the rest of this documentation, you’ll learn how to:

* How to setup Governed Metrics
* How to create and manage metrics
* Data cleaning and common pitfalls

<br>