Customizing Chat Results

Customizing Chat-Driven Analytics with Grounding, Joins, and Follow Ups

Zing’s chat interface let you ask questions in natural language and get charts and text summaries in seconds. But you can also control and customize this experience with examples, relationships, and definitions to further refine the answers that Zing provides.

These work out of the box

• Automatic correction of misspelled terms (e.g., “Califronia” to “California”)
• Field detection without explicit column naming (automatically maps state names to state field)
• Contextual understanding for follow-up questions (e.g., “add Arizona” or “only the top three”)
• Dynamic switching between chart types (bar, stacked bar, donut)
• Real-time alerts based on updated data
• Automated key point summarization above visualizations
• Visual editor for fine-tuning chart appearance and data display
• Automatic detection and execution of necessary table joins
• Transparent join relationship visibility (e.g., inner join on order ID) so you always see what Zing is doing
• Full visibility into underlying SQL queries
• Access to calculation formulas and join logic
• Three ways to ask questions and view results : chat, visual editor, and SQL IDE

Further tuning

You can further tune of Zing’s chat experience responds by defining example questions, aliases, and calculations, as demonstrated in the video above.

This lets you:

• Define business metrics (for example, ROAS - Return on Ad Spend, or days of inventory) by adding examples

• Provide examples for complex or ambiguous queries and how you want Zing to resolve them (e.g. define an engaged user as more than x minutes on the app)

• Create a metric, alias, or example that Zing can apply to different dimensions

   

Get started with Zing for free! Sign In or create a New Account.