Skip to main content

AI Chat Interface

Your primary way to work with data—just type what you want in plain English.

What is the AI Chat Interface?​

The chat box at the bottom of your Workbook is where all the magic happens. Instead of writing SQL or clicking through menus, you simply type requests in natural language. The AI understands your intent, generates Views, creates visualizations, and answers questions—all from conversation.

AI Chat Interface

The chat interface is always available at the bottom of your Workbook

Why the Chat Interface Matters​

Natural interaction: Talk to your data like you'd talk to a data analyst.

Context-aware: The AI remembers your previous questions and builds on prior work.

Flexible: Ask questions, request transformations, create charts, or explore—all in one place.

Transparent: See exactly what the AI creates. Nothing hidden.

How to Use the Chat Interface​

Basic Requests​

Simply describe what you want:

Show me total revenue by month
Find customers who haven't purchased in 90 days
Calculate the average order value for each product category

The AI will create Views, process the data, and show results.

Referencing Data with @Mentions​

Use @ to reference specific data objects:

@[orders] Show the top 10 products by revenue
Join @[customers] with @[orders] to show customer lifetime value

As you type @, autocomplete suggests available Sources, Views, and Input Objects.

@ Autocomplete

Type @ to see all available data objects

Building on Previous Work​

The AI remembers context. You can reference earlier Views naturally:

Using that revenue analysis, show me which regions are growing fastest
Now filter to only Q4 results
Group that by customer segment

Conversational Flow​

Unlimited Exchanges​

Chat as much as you need. Ask follow-ups, request modifications, dig deeper:

You: Show monthly sales trends AI: Creates Views and chart You: Now break it down by region AI: Adds regional dimension You: Filter to only regions with growth over 10% AI: Applies filter

When to Start Fresh​

After ~10 exchanges, consider starting a new chat to keep context focused. Click "Start new chat" to clear the interface while keeping all your Views and data.

Your previous chat history is always accessible via the history icon.

Chat History

Access previous conversations anytime

Common Use Cases​

Exploratory analysis: "Show me interesting patterns in customer behavior"

Specific calculations: "Calculate year-over-year growth for each product"

Data cleaning: "Remove rows where email is null or invalid"

Joins and combinations: "Combine sales data with product info"

Aggregations: "Give me monthly averages, grouped by category"

Filtering: "Show only orders above $500 from this year"

Time-based analysis: "Compare Q4 2023 vs Q4 2024 performance"

Tips & Best Practices​

Be specific: "Show monthly revenue for 2024" is clearer than "Show revenue"

Use @mentions: Always reference data explicitly with @[data_name] for best results

Ask iteratively: Start broad, then refine. Don't try to describe everything in one message.

Request explanations: Ask "Why did you do it that way?" or "Explain this View" if unclear

Verify assumptions: Use /plan mode (see Slash Commands) to see the AI's approach before it executes

Check intermediate steps: Look at the Views created, not just the final result

Be patient: Complex requests may take a few seconds as the AI generates multiple Views

Understanding AI Responses​

The AI can:

  • Create new Views and show data
  • Generate visualizations
  • Explain its reasoning
  • Suggest next steps or related analyses
  • Point out potential issues in data or logic
  • Recommend optimizations

If the AI asks clarifying questions, answer them to help it understand your intent better.

AI Response

AI explanations alongside generated Views

Special Features​

Auto-save: Every message is automatically saved—never lose your work

Copy-paste friendly: Paste SQL, example data, or specifications directly into chat

Multi-line input: Shift+Enter for line breaks within your message

Rich formatting: The AI's responses include formatted code, tables, and structured explanations