> 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/what-can-and-cannot-ai-do.md).

# What can and cannot AI do?

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

### TL;DR

The AI assistant automates data exploration and content generation but requires manual verification and does not perform non-data related tasks.

***

### When to use this

Use this page to understand the functional boundaries of the AI assistant when generating queries, charts, and metadata.

***

### What Toucan AI Can Do

The AI assistant performs the following tasks within the Toucan.ai environment:

* **Schema Interpretation**: Identifies relevant tables and columns based on the connected database structure.
* **Query Execution**: Generates and executes SQL queries to answer natural language questions.
* **Visualization Generation**: Recommends and builds chart types based on the identified data and user intent.
* **Dashboard Assembly**: Constructs complete dashboard layouts from a single prompt.
* **Contextual Explanation**: Summarizes data findings and explains the logic used to create a visualization.
* **Metadata Enrichment**: Populates descriptions and semantic field types for database tables and columns.

***

### What Toucan AI Cannot Do

The AI assistant is subject to the following technical and operational limits:

* **No Image Generation**: The AI assistant only produces text-based insights and data visualizations.
* **No Data Fabrication**: Toucan.ai does not guess or synthesize data; it only visualizes results from executed queries.
* **No Direct Config Manipulation**: The AI assistant cannot modify the granular configuration of a chart once it is placed in a dashboard; manual editing is required for these changes.
* **Data-Only Scope**: The AI assistant does not respond to casual conversation or non-data prompts.
* **Manual Review Required**: The AI assistant may produce incorrect inferences; users must validate all metadata and chart outputs before production.
* **Security Adherence**: The AI assistant cannot bypass defined Row-Level Security (RLS) or user permissions.