AI Data Analyst
QRY's AI Data Analyst understands natural language questions and automatically generates SQL queries to answer them.
How It Works
- You Ask: Type your question in plain English
- AI Understands: The AI analyzes your question and database schema
- SQL Generated: QRY writes the SQL query automatically
- Results Delivered: Data is retrieved and displayed in an easy-to-read format
Supported AI Models
QRY supports multiple cutting-edge AI providers and models:
Claude (Anthropic)
- Claude Sonnet 4.5: Best for complex analytical queries with extended context
- Claude Haiku 4.5: Lightning-fast responses for simpler queries
- Claude Opus 4.1: Maximum capability for the most challenging analysis
Gemini (Google)
- Gemini 2.5 Pro: Advanced reasoning with enhanced multimodal capabilities
- Gemini 2.5 Flash: Quick responses with excellent accuracy and efficiency
GPT (OpenAI)
- GPT-5: State-of-the-art language understanding and generation
- GPT-4 Turbo: Proven performance with faster response times
Model Selection
Your administrator controls which models are available to you. Default models can be set per user group.
Features
Natural Language Processing
Ask questions the way you'd ask a colleague:
What were our top 5 products by revenue last month?
Show me customer retention rate by quarter
Compare sales across regions for the past year
Automatic SQL Generation
QRY translates your questions into optimized SQL queries. You can:
- View the generated SQL
- Edit the SQL if needed
- Learn SQL by seeing how questions translate to queries
Context Awareness
QRY remembers your conversation:
- Follow-up questions don't need full context
- Reference previous results: "Now show me the same for last year"
- Build complex analysis step by step
Schema Understanding
The AI automatically:
- Discovers table relationships
- Identifies relevant columns
- Suggests appropriate joins
- Handles date/time formats
Best Practices
Writing Effective Questions
Do:
- Be specific about time periods
- Mention exact metrics you want
- Include filtering criteria
- Request specific aggregations
Don't:
- Use ambiguous terms like "recently" or "some"
- Mix multiple unrelated questions
- Assume the AI knows unstated context
Example Questions
Sales Analysis:
Calculate total revenue and average order value by product category for Q3 2024
Customer Insights:
Find customers who purchased more than 3 times in the last 6 months but haven't bought anything in the last 30 days
Trend Analysis:
Show monthly active users for the past 12 months with percentage change from previous month
Comparisons:
Compare year-over-year revenue growth for each sales region
Advanced Features
Multi-Step Analysis
Break complex questions into steps:
- "Show me top 10 customers by revenue"
- "What products did they buy most frequently?"
- "Create a chart showing their purchase patterns over time"
Data Exploration
Discover your data:
Describe the structure of the orders table
What are the unique values in the status column?
Show me sample records from the customers table
Visualizations
Request charts and graphs:
Plot monthly sales as a line chart
Create a bar chart comparing product categories
Show customer distribution by region on a pie chart
Limitations
While QRY's AI is powerful, keep in mind:
- Database Permissions: You can only query data you have access to
- Query Complexity: Extremely complex queries may require breaking into steps
- Real-time Data: Results reflect data available at query time
- Token Limits: Very large result sets may be truncated
Tips for Success
- Start Broad, Then Refine: Begin with general questions, then add filters and specifics
- Review Generated SQL: Learn from the AI's query construction
- Use Follow-ups: Build on previous answers for deeper insights
- Save Good Queries: Bookmark conversations with useful analysis for reference
- Experiment: Try different phrasings to see what works best
Learn More
Check out the Quick Start Guide for hands-on examples and the User Guide for advanced techniques.