> For the complete documentation index, see [llms.txt](https://docs.toucanai.cloud/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.toucanai.cloud/faq-and-troubleshooting/ai-assisted-creation/where-is-ai-used-in-toucan.md).

# Where is AI used in Toucan?

{% hint style="info" %}
**Target Audience**: Non technical users & Developers
{% endhint %}

### TL;DR

Toucan I integrates Large Language Models (LLMs) across the platform to automate metadata generation, query construction, and visualization.

***

### When to use this

Use this page to identify which parts of the analytics workflow support automation and where the AI assistant replaces manual configuration.

***

### AI Integration Points

Toucan AI utilizes artificial intelligence at multiple stages of the analytics lifecycle to process data schemas and generate content.

**1. Automated Metadata Enrichment**

* **Contextual Analysis**: Toucan AI scans database schemas to generate table and column descriptions.
* **Semantic Inference**: Toucan AI identifies field types (Category, Date, Metric) to prepare data for visualization.

**2. Chart and Dashboard Generation**

* **Natural Language Processing**: The AI assistant converts text prompts into SQL queries and visualizations.
* **Visual Selection**: The AI assistant identifies the chart type best suited for the requested metrics and dimensions.
* **Step Explanation**: Toucan AI provides a technical summary of how the query was constructed and the logic applied to the data.

**3. Conversational Exploration**

* **Assistant Interface**: The AI assistant is available in the Library and on dashboards for ad-hoc data questions.
* **Pattern Detection**: The AI assistant identifies statistical trends within tables and suggests relevant dimensions for further exploration.

**4. Embedded Conversational Analytics**

* **End-User Access**: The AI assistant can be integrated into host applications via web components.
* **Self-Service**: End-users can generate custom charts and queries using the AI assistant within the embedded environment.

***

### Technical Architecture

Toucan AI utilizes a multi-layered approach to interact with data securely.

* **Execution Model**: AI-generated visualizations are based on real-time query execution against your connected database.
* **Data Security**: Toucan AI interacts with schema metadata and sampled values to build prompts; it does not move entire datasets into the LLM.
* **Determinism**: All generated charts are derived from actual data values rather than synthetic or guessed samples.

***

### Constraints

* **Validation**: All AI-generated metadata and charts should be reviewed for accuracy before production deployment.
* **Schema Dependence**: The quality of the AI assistant output is directly proportional to the clarity of the defined metadata.