Using Data Conversations

Using Data Conversations#

This guide explains how to use Data Conversations, Adverity’s natural language interface for exploring your data.

Note

For customers who require their data to be processed and stored exclusively within the European Union (EU), Adverity uses Microsoft’s Azure OpenAI Service. This service is hosted in the EU and complies with all EU data residency requirements.

Introduction#

What are Data Conversations?

Data Conversations reshape how you interact with your data by allowing you to simply ask questions and receive immediate insights. This product enables anyone in your organization to access data insights without technical expertise or SQL knowledge.

Whether you need quick answers about campaign performance, want to identify trends, or need to share insights with stakeholders, Data Conversations makes your data truly accessible and actionable.

What are the benefits of using Data Conversations?

With Data Conversations you can:

  • Get immediate answers about your data without writing SQL queries.

  • Empower team members of all technical levels to explore data independently.

  • Reduce time spent on repetitive data queries and report creation.

  • Access the information you need when you need it for better decision-making.

  • Easily share data insights with colleagues and clients.

How do Data Conversations use AI?

Data Conversations uses large language models (LLMs) to power its conversational AI capabilities. Adverity leverages OpenAI technology for processing natural language requests, specifically the GPT-5 model, starting from release 2026.06 (GPT-4.1 was used previously). Only column names and necessary database metadata are sent to OpenAI to generate SQL queries, which are then executed within Adverity’s secure environment. Your actual data remains within Adverity, and no personal data is shared with external providers. The summary statistics of your data may be shared with the LLM if it is requested in your prompt, however, the dataset itself remains in Adverity.

What Data Conversations response may look like?

Data Conversations answer may include:

  • Text explanations of the data

  • Data tables with relevant information

  • Visualizations (bar or line charts)

Additionally, you can view an SQL query used to get the data relevant to your prompt.

The response’s data size is limited to 10,000 rows per query.

Can I access my previous conversations?

All your conversations are saved automatically and can be opened from the Your conversations section. You can rename and delete saved conversations. Additionally, you can add the selected conversations to favorites for easy access.

To organize and share the insights from Data Conversations, add your findings to notebooks. For more information, see Working with Data Conversations notebooks.

Prerequisites#

Before using Data Conversations, perform all of the following actions:

Hint

If you transform your data outside of Adverity, set up a datastream in a new workspace to load it into Adverity Data Storage or your own Snowflake warehouse connected to Adverity after it has been transformed. This way you can get Data Conversations insights after the full processing.

Data Conversations best practices#

To get the most relevant insights with Data Conversations, use our best practices:

Filtering data#

You can apply filters to narrow down the data used in your conversations. Below the prompt text box, select the workspaces and datastreams from which the data will be used.

To apply consistent filters, use quick filters. Quick filters are pre-configured filter sets that can be saved and reused in Data Conversations. Instead of manually setting up filters every time you want to analyze data, you can create a quick filter once and apply it with a single click.

Key benefits of quick filters:

  • Quick filters are accessible by all users working with Data Conversations.

  • Set up complex filters once and reuse them indefinitely.

  • Ensure consistent filtering logic across your team for comparable results.

Adding a quick filter#

To add a quick filter, follow these steps:

  1. Below the prompt text box, click Quick filter.

  2. Click Add quick filter.

  3. Enter a name for the quick filter.

  4. Click Add quick filter.

Using quick filters#

To apply a quick filter to your prompt, select the filter from the list of available quick filters and click Apply.

Existing quick filters can be edited or deleted.

Working with responses#

After getting a response, you’ll see suggested follow-up questions. Use these to dig deeper into your data or explore related aspects.

To understand what data was used to create the insights, click View source under any response. This shows you a summary of datastreams from which the selected fields are being pulled as well as the SQL query executed on your warehouse.

To add a conversation to favorites, open the image1 More menu for the conversation and click image2 Add to favorites.

Favorite conversations are displayed on top of your conversation history and are easy to access.

For data tables

Click Download data image3 above the insight to save it as a CSV file.

For charts and visualizations

Click Download image image3 above the insight to save it as an image file.

To export your findings in the table format to Google Sheets for external use, follow these steps:

  1. Click Export to Google Sheets above the response.

  2. Sign in to your Google account or select an already authorized account.

  3. The created export file will be linked from your conversation.

To add an insight to a notebook, follow these steps:

  1. Click Add to notebook below the response.

  2. Select the target notebook from the dropdown menu.

The insight will be saved to the selected notebook with the original date and source conversation information. For more information, see Working with Data Conversations notebooks.

Contents