Overview

The AI agent understands natural language questions about your data and automatically creates visualizations and analyses. Learn the main categories of questions that work best.

The AI agent processes your questions in natural language and:

• Interprets Intent: Understands what type of analysis you want
• Identifies Variables: Determines which columns to analyze
• Selects Methods: Chooses appropriate analytical techniques
• Creates Visualizations: Generates charts that answer your question

Once you ask a question, the agent:

1. Analyzes your data to understand structure and relationships
2. Applies appropriate techniques for the question type
3. Creates relevant visualizations to show findings
4. Provides explanations of what the results mean

❌ Vague: “Analyze my sales data” ✅ Specific: “What factors drive sales performance in the North region?”

❌ No Context: “Why did this number change?” ✅ With Context: “Why did conversion rates drop after we changed the pricing page?”

Start broad, then drill down:

1. “What are the main patterns in customer behavior?”
2. “Why do customers in segment A behave differently?”
3. “What would happen if we targeted segment A with personalized campaigns?”

• Start with exploratory questions to understand your data
• Follow up with specific questions about interesting patterns
• Drill down into areas that show potential insights
• Ask comparative questions to validate findings

The agent can handle sophisticated questions that require multiple analytical steps:

Example: “Why did our Q3 revenue miss the target, and what factors should we focus on?”

Agent’s Approach:

1. Analyzes Q3 revenue vs. target
2. Identifies contributing factors through correlation analysis
3. Compares to historical patterns
4. Creates visualizations showing key relationships
5. Highlights areas for further investigation

• Root Cause Analysis: “Why did X happen and what can we do about it?”
• Impact Assessment: “What would happen if we change Y?”
• Opportunity Identification: “Where are the biggest opportunities for improvement?”
• Strategy Validation: “Is our hypothesis about Z supported by the data?”

• Specific objectives: What you want to understand or decide
• Clear scope: Which data or time period to focus on
• Context: Why this analysis matters for your goals
• Success criteria: How you’ll know if the analysis is useful

• Overly broad requests without direction
• Questions that assume conclusions
• Requests for analysis without clear purpose
• Questions that require data you don’t have