Understand how AI works with your data.
Bold BI's AI features help you create visualizations, ask questions about your data, and summarize dashboard information. These capabilities depend on the features you use, the AI provider configuration, and the access controls applicable to your deployment. Bold BI does not use customer data to train or fine-tune its own AI models.
Features and administrative controls vary by product version, deployment, and configuration.
Each principle explains what Bold BI's AI features do with your information and what depends on your provider and configuration.
Bold BI does not use customer data to train or fine-tune its own AI models.
What this means
The information included in an AI request depends on the task.
What this means
AI features use Bold BI’s authentication, authorization and data-source security framework.
What this means
Bold BI supports a built-in AI service and configurations that use your own provider credentials.
What this means
AI-generated answers, summaries and visualizations can contain errors or omit relevant context.
What this means
For deployments using Bold BI’s managed AI service, applicable credit and rate-limit controls govern usage.
What this means
The information included in an AI request depends on the task. Questions, metadata, and query results can contain sensitive information. Consider the information you include in a prompt and the provider configured to process it.
| AI task | Information that may be processed |
|---|---|
| Create or update a visualization | Your question, data source schema, configured value synonyms, and relevant instructions or conversation context. |
| Answer a question about data | Your question, relevant data source metadata, query results needed to formulate the answer, and relevant context or instructions. |
| Summarize a widget or dashboard | Your question where applicable, widget titles and types, column bindings, displayed values or data points, and relevant context. |
| Use an AI Assistant or connected tool, where available | Information needed for the task, including relevant conversation context, instructions, and applicable tool inputs or results. Connected services may introduce additional processing destinations. |
Query results may contain aggregated values or detailed records, depending on the query and applicable access controls. These results should not be assumed to be anonymous.
Features and administrative controls vary by product version, deployment, and configuration.
| Control | What it does | Applies to |
|---|---|---|
| AI provider choice | Use Bold BI’s built-in AI service, or configure your own provider credentials for OpenAI, Azure OpenAI or Anthropic. | Subject to your product version and deployment. |
| Access controls | AI features use Bold BI’s authentication, authorization and data-source security, including row-level security and embedded-user identity where configured. | Depends on the user’s identity, permissions, data source configuration and the operation being performed. |
| Managed-service credits and rate limits | Govern AI usage; each request is checked against the available credit balance. | Deployments using Bold BI’s managed AI service. |
| Provider billing and limits | Usage billing and account-level limits are managed through your provider account. | Deployments using your own provider credentials. |
| Response feedback | Rate responses to help the Bold BI team identify issues and assess improvements. | Where feedback controls are available. |
Review the AI configuration documentation for the options available in your deployment.
Depending on the feature, deployment, and configuration, Bold BI may store conversation history, request and response information, usage records, and feedback. Removing a conversation from the history view does not necessarily erase all associated stored records.
Storage, access, retention, and deletion depend on the deployment and service involved. Review the applicable privacy and service documentation, or contact Bold BI support for details. Provider-side storage and retention are governed separately by the provider's terms and configuration.
AI-generated answers, summaries, and visualizations can contain errors or omit relevant context. Review the underlying data, filters, calculations, and generated output before relying on them for important decisions.
If an AI response appears incorrect or exposes information unexpectedly, report it to Bold BI support.
Bold BI does not use customer data to train or fine-tune its own AI models. AI requests are processed by the provider configured for your deployment, and when you use your own provider credentials, that provider's terms, account settings and data-processing policies also apply. Review the provider's documentation for its training, monitoring, storage and retention practices.
It depends on the task. An AI request may include your question, data source metadata such as the schema and configured value synonyms, query results, widget data, and relevant conversation context or instructions. Query results may contain aggregated values or detailed records, depending on the query and applicable access controls, and should not be assumed to be anonymous.
Bold BI supports a built-in AI service and configurations that use your own provider credentials. Supported provider integrations include OpenAI, Azure OpenAI and Anthropic, subject to your product version and deployment, and the available models depend on the selected provider and configuration.
Yes. When you use your own credentials, the provider's terms, account settings and data-processing policies apply, and provider usage billing and account-level limits are managed through your provider account. These are separate from the credit and rate-limit controls of Bold BI's managed AI service.
AI features use Bold BI's authentication, authorization and data-source security framework, so access depends on the requesting user's identity, permissions, data source configuration and the operation being performed. Administrators should configure and validate the security rules required for their deployment, including row-level security and embedded-user identity where applicable.
Depending on the feature, deployment and configuration, Bold BI may store conversation history, request and response information, usage records and feedback. Removing a conversation from the history view does not necessarily erase all associated stored records. Storage, access, retention and deletion depend on the deployment and service involved; contact Bold BI support for details.
AI-generated answers, summaries and visualizations can contain errors or omit relevant context. Review the underlying data, filters, calculations and generated output before relying on them for important decisions, and refine generated visualizations using the dashboard design tools.
For deployments using Bold BI's managed AI service, applicable credit and rate-limit controls govern usage, and requests are checked against the available credit balance. When you use your own provider credentials, billing and limits are set by your provider account instead.
Review the AI configuration documentation, security information, privacy policy, and applicable provider data-processing documentation. Contact Bold BI support with questions about your deployment or to report an AI response that appears incorrect or exposes information unexpectedly.
Ask questions in plain English, then review and refine every answer and visualization before you act on it.
