Skip to main content

Querying Your Data

Once your Sphere is configured, you can start asking questions using the chat interface.

The Chat Interface​

The chat interface is where you interact with your Sphere. Type your questions in natural language, and the AI will translate them into SQL, query your database, and return the results.

Conversation Starters​

The yellow boxes above the input field are Conversation Starters. These are pre-configured example questions designed to help you understand the types of queries your Sphere can handle. Click on one to automatically send it.

The DeepSphere Toggle​

Below the text input, you will find the DeepSphere toggle. This feature controls the depth of the AI's analysis.

  • Disabled (Default Mode): This mode is optimized for fast, direct answers to specific questions. It's best for straightforward queries like "How many users signed up last month?".

  • Enabled (Deep Analysis Mode): When enabled, DeepSphere performs a more comprehensive, multi-step analysis. It might break down your question into several sub-queries, analyze the results, and provide a more detailed narrative or comparative insight. This mode is ideal for exploratory questions like "Analyze the sales trends of our top 3 product categories over the last year."

    Credit Consumption

    Enabling DeepSphere will consume more credits per query, as it involves more complex reasoning and potentially multiple interactions with the AI model.

Asking Effective Questions​

The quality of your input directly affects the quality of the output. Here are some tips for writing effective queries:

  • Be specific: Instead of "show me sales," ask "What were the total sales in USD for the month of October 2025?".
  • Define the scope: If you want a comparison, state it clearly. "Compare the number of new users from 'Google' vs. 'Facebook' traffic sources."
  • Use column names: Although optional, if you know the names of your tables and columns, using them can help the AI generate more accurate queries. "What is the average order_value from the orders table?"