> For the complete documentation index, see [llms.txt](https://docs.datalogz.io/guides/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.datalogz.io/guides/semantic-intelligence/how-semantic-intelligence-works.md).

# How Semantic Intelligence works

[Semantic Intelligence](https://docs.datalogz.io/guides/semantic-intelligence.md) runs in five steps for each connector.

1. **Read.** Control Tower collects every metric from your semantic models (DAX measures and calculated columns, calculation fields, metric views, etc), with its name, table, data model, department, certification and usage. Excluded data models are skipped.
2. **Parse.** Each expression is expanded (referenced metrics are written out in full) and parsed into a tool-agnostic logical tree. Formatting, variable names and equivalent ways of writing the same logic disappear at this step.
3. **Compare.** Promising pairs of definitions are scored on four signals: Structure, meaning, linguistics, metadata.
4. **Group.** Definitions that score high enough are grouped into a business concept, and a source definition is chosen for each concept. Each candidate must also be close enough to the source definition itself to stay in the concept; candidates that are not are split off and grouped again.
5. **Classify.** Every definition is labeled against the reference definition, and each concept gets its status.

#### How the reference definition is chosen

Within each concept, Control Tower selects the metric that is most likely to be the trusted version. It favors:

* metrics that other semantic models reference as their source
* metrics in a certified (endorsed) semantic model
* metrics used by more reports
* metrics most similar to the rest of the concept

#### When results update

Semantic Intelligence analyzes the latest metadata from your connector scans. Results are scoped to each team's workspaces, so two teams on the same connector each see their own concepts. While an analysis run is in progress, previous results stay visible until the new run is published.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.datalogz.io/guides/semantic-intelligence/how-semantic-intelligence-works.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
