> ## Documentation Index
> Fetch the complete documentation index at: https://docs.chartcastr.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Conversational data analysis

> Ask complex questions about your data in Chartcastr AI chat. The assistant uses external context, pulse history, and documents for deeper insights.

## Overview

Chartcastr AI Chat provides a conversational interface for deep analysis of your data. It goes beyond simple snapshots, allowing you to ask complex questions that leverage the full [Context](/analysis/context/overview) of your workspace.

### Context Discovery

The AI automatically discovers context from your [connected sources and documents](/analysis/context/overview) to make every interaction smarter. It doesn't just look at the numbers; it reads your strategy docs and metric definitions to explain *why* the numbers are changing.

## Context-Aware Analysis

The power of Chat analysis comes from its ability to synthesize multiple information sources:

### Pulse and Connection Data

Chat has direct access to your latest pulse data and connection history.

* *"Why did revenue drop yesterday compared to the same day last week?"*
* *"Which items are currently trending in our inventory?"*

### External Context Integration

If you've linked [External Context](/analysis/context/overview) (like Google Docs or Sheets), Chat incorporates that information into its analysis.

* *"Based on our Q1 Strategy document, how are we tracking against our growth goals?"*
* *"Explain the latest revenue numbers in the context of our pricing update definitions in Sheets."*

### Comparing Data

You can use Chat to compare performance across different sources or time periods.

* *"Compare the conversion rate of our Facebook ads vs Google ads over the last 30 days."*
* *"Show me how the new product launch in London compares to the Manchester launch."*

## Interactive Exploration

Analysis in Chat is iterative. You can start with a broad question and drill down into specific details.

1. **Start Broad**: *"How was our performance last week?"*
2. **Drill Down**: *"Break down that revenue spike by category."*
3. **Analyze Root Cause**: *"Was that increase driven by new customers or existing ones?"*
4. **Get Recommendations**: *"What should we focus on next week to maintain this momentum?"*

## Visualizing Insights

While Chat is primarily text-based, it can help you prepare data for visualization.

* *"Summarize these findings into a list of bullet points for my weekly report."*
* *"Format the revenue comparison as a table."*

## Privacy and Data Security

Chat analysis respects all your existing permissions and data security settings. It only analyzes data that you have explicit access to within the workspace.