Beginner guide to natural language queries, question formatting, AI interpretation, and getting the most accurate results.
What Are Natural Language Queries?
Overview
Natural language queries let you ask questions about your data using everyday English instead of complex SQL or programming languages. clariBI's AI understands your questions and automatically generates the appropriate analysis.
Benefits
- No Technical Skills Required: Ask questions in plain English
- Instant Insights: Get immediate answers to business questions
- Exploratory Analysis: Discover patterns through conversation
- Time Savings: No need to build custom reports
- Accessible to Everyone: Democratizes data analysis
Getting Started
Accessing AI Chat
- Dashboard View: Click the AI Chat button in any dashboard
- Analytics Page: Navigate to Conversational Analytics
- Quick Access: Use the AI icon in the navigation bar
- Widget Creation: Select "Natural Language" when adding widgets
Your First Query
Try these simple starter questions:
- "What were our sales last month?"
- "Show me top performing products"
- "How many new customers did we get this week?"
- "What's our conversion rate trend?"
Effective Query Techniques
Question Structure
```
Good Query Format:
[What/Show me/How many] + [metric] + [time period] + [optional filters]
Examples:
• "What were our website conversions last quarter?"
• "Show me revenue by product category this year"
• "How many support tickets were closed yesterday?"
• "What's our customer churn rate by region?"
```
Specific vs. General
```
Instead of: "Show me sales"
Try: "Show me monthly sales for the last 6 months"
Instead of: "Customer data"
Try: "What's our customer acquisition rate this quarter?"
Instead of: "Marketing performance"
Try: "Which marketing channels generated the most leads last month?"
```
Types of Questions You Can Ask
Basic Metrics
- "What's our total revenue this month?"
- "How many users signed up today?"
- "What's our current inventory level?"
- "Show me website traffic this week"
Comparisons
- "Compare sales this month vs last month"
- "How does revenue this quarter compare to same quarter last year?"
- "Which product category performed better: electronics or clothing?"
- "Compare conversion rates between mobile and desktop"
Trends and Patterns
- "Show me sales trend over the last 12 months"
- "Is our customer acquisition rate increasing?"
- "What's the pattern in website traffic by day of week?"
- "How has our churn rate changed over time?"
Top/Bottom Lists
- "Who are our top 10 customers by revenue?"
- "What are our worst performing products?"
- "Which sales reps have the highest conversion rates?"
- "Show me top traffic sources this month"
Filtering and Segmentation
- "Sales in California last month"
- "New customers from organic search"
- "Support tickets marked as urgent"
- "Revenue from enterprise customers only"
Advanced Query Examples
Multi-Dimensional Analysis
```
• "Show me revenue by product and region for Q4"
• "Customer acquisition cost by marketing channel and month"
• "Support ticket resolution time by priority and team"
• "Website conversion rate by traffic source and device type"
```
Time-Based Queries
```
• "Year-over-year growth rate for the last 3 years"
• "Monthly recurring revenue trend since January"
• "Daily active users for the past 30 days"
• "Quarterly sales performance compared to targets"
```
Conditional Logic
```
• "Customers who made more than 3 purchases this year"
• "Products with inventory below 50 units"
• "Support tickets open for more than 7 days"
• "Users who haven't logged in for 30 days"
```
Improving Query Accuracy
Be Specific About Time
- Instead of "recent": use "last 30 days", "this month", "Q3 2024"
- Specify time zones when relevant
- Use consistent date formats
- Be clear about business vs calendar periods
Use Exact Field Names
- Learn your data field names: "customer_id" vs "user_id"
- Use consistent terminology: "revenue" vs "sales" vs "income"
- Specify units when needed: "revenue in dollars" or "weight in pounds"
- Be precise about categories: "product_category" vs "product_type"
Provide Context
```
Better Queries with Context:
• "Show me B2B customer churn rate for enterprise accounts"
• "Website conversion rate for organic traffic from Google"
• "Average order value for repeat customers in the US"
• "Support ticket resolution time for technical issues"
```
Understanding AI Responses
Response Types
- Charts and Visualizations: Graphs showing trends and comparisons
- Tables: Detailed data listings and rankings
- Key Metrics: Single number answers with context
- Insights: AI-generated observations and recommendations
- Follow-up Suggestions: Related questions you might ask
When AI Needs Clarification
The AI might ask for clarification like:
- "Which time period did you mean?"
- "Did you want to see this by product category or individual product?"
- "Should I include refunded orders?"
- "Do you want to filter by any specific region?"
Common Challenges and Solutions
Query Not Understood
Problem: AI doesn't understand your question
Solutions:
- Rephrase using simpler language
- Break complex questions into parts
- Use field names from your data
- Try alternative synonyms
Unexpected Results
Problem: Results don't match expectations
Solutions:
- Check date ranges and filters
- Verify data source is correct
- Ask follow-up clarifying questions
- Review underlying data quality
Missing Data
Problem: AI says no data found
Solutions:
- Adjust time periods
- Check spelling of field names
- Verify data source connections
- Try broader search terms
Best Practices
Query Strategy
- Start Simple: Begin with basic questions
- Build Complexity: Add filters and dimensions gradually
- Validate Results: Cross-check important findings
- Save Useful Queries: Bookmark frequently used questions
- Share Insights: Export and share interesting discoveries
Data Preparation
- Clean Data Sources: Ensure data quality
- Consistent Naming: Use standard field names
- Regular Updates: Keep data sources current
- Document Definitions: Maintain data dictionary
- Test Connections: Verify data source reliability
Team Adoption
- Training Sessions: Teach team members effective querying
- Example Library: Create common query examples
- Best Practices: Share successful query patterns
- Regular Review: Assess and improve query techniques
- Feedback Loop: Collect and act on user feedback
Advanced Features
Follow-up Questions
After getting initial results, ask:
- "Why did sales drop in March?"
- "What caused this spike in traffic?"
- "Can you break this down by region?"
- "Show me the same data for last year"
Automated Insights
Enable automatic insights to:
- Detect anomalies in your data
- Identify significant trends
- Suggest relevant questions
- Provide contextual recommendations
Query History
- Save Queries: Bookmark useful questions
- Query Templates: Create reusable question formats
- Sharing: Share successful queries with team
- Scheduling: Set up automated query reports
Natural language queries make data analysis accessible to everyone, regardless of technical background. Start with simple questions and gradually build your skills to unlock deeper insights from your data.