How It Works
When you upload a data file or ask analytical questions, the AI has access to 15 specialized tools for data processing. It automatically determines which tools to use based on your request. For example, asking “Show me the revenue trend over the last 12 months” triggers the AI to:- Read the relevant data file
- Identify the date and revenue columns
- Run a time series analysis
- Generate a line chart
- Present the chart with a written summary of the trend
Analysis Focus Modes
When the AI analyzes your data, it selects a focus mode based on what you are asking:| Mode | What It Does | Best For |
|---|---|---|
| Comprehensive | Full analysis across multiple dimensions with several chart types | ”Give me an overview of this dataset” |
| Distribution | Histograms, box plots, and statistical summaries of how data is spread | ”How is salary distributed across the company?” |
| Correlation | Scatter plots and heatmaps showing relationships between variables | ”Is there a relationship between ad spend and conversions?” |
| Time Series | Line charts and trend analysis over temporal data | ”Show me monthly revenue for the past year” |
| Comparison | Bar charts and ranking visualizations for categorical data | ”Compare sales performance across regions” |
Chart Types
The AI generates publication-ready charts directly in the conversation:Bar & Stacked Bar
Categorical comparisons and composition breakdowns
Line
Trends over time and sequential data
Scatter
Relationships between two numeric variables
Histogram & Box
Data distribution and statistical spread
Pie
Proportional composition of a whole
Heatmap & Ranking
Correlation matrices and ordered comparisons
Explore and Compute Tools
Beyond full analysis, the AI can perform quick data operations: Explore Data lets the AI preview your dataset before running a full analysis. It supports actions like:- Preview — See the first few rows and column types
- Describe — Get statistical summaries (mean, median, min, max, standard deviation) for every numeric column
- Columns — List all columns with their data types
- Sum, Mean, Median — Aggregate a column
- Group By — Break down metrics by category
- Rank — Order rows by a specific column
- Pivot — Restructure data into a pivot table
- Correlation — Calculate correlation coefficients between columns
Inline Data Analysis
You do not need an uploaded file to generate charts. The AI can visualize data from any source that appears in the conversation, including:- Data returned from company knowledge base queries
- Numbers you type directly into the chat
- Tables extracted from documents discussed earlier in the conversation
Querying Spreadsheets with Natural Language
Uploaded CSV and Excel files can be queried using plain English. Ask questions like:- “What is the total revenue for Q3?”
- “Show me all rows where the status is ‘overdue’”
- “What is the average deal size by sales rep?”
Natural language queries work on any CSV or Excel file in Ready status. The AI reads the column headers to understand the data structure, so descriptive column names produce better results than generic ones like “Column A”.
Tips for Better Analysis
- Be specific about what you want to see — “Show revenue by quarter” gives better results than “Analyze this file”
- Name your columns clearly — AI performance improves when spreadsheet headers are descriptive
- Ask follow-up questions — After an initial analysis, drill deeper with questions like “Now break that down by region” or “Remove the outliers and re-chart”
- Request specific chart types — If you want a scatter plot instead of the bar chart the AI chose, just ask: “Show that as a scatter plot instead”