Create visuals from a prompt
Describe a chart and AI Copilot builds it. Attach a datasource and it suggests 4 sample queries.
AI Copilot · Dashboard Designer
Bold BI AI Copilot builds charts from plain-English requests inside the Dashboard Designer. Tell it what you want, and it builds the widget from your connected datasource and picks a fitting chart type. Refine it by asking, since Copilot considers your previous 4 conversations, or by hand. It runs on Cloud and on-premises, including Docker and Kubernetes.
No credit card required

How it works
AI Copilot works in 3 steps inside the Dashboard Designer: ask in plain English, it builds the widget, then refine it.
Type what you want in plain English: a chart, a comparison, a KPI. Type @ to pick one datasource or several.
It builds the widget from your connected data and picks a fitting chart type.
Refine by follow-up prompt, since Copilot considers your previous 4 conversations, or edit the widget by hand.
Want a whole dashboard from one sentence — not one chart at a time? That's Prompt to Dashboard, which builds a starter dashboard of 6 widgets. Copilot is for building and refining individual visuals.
Capabilities
AI Copilot creates visuals from a prompt, edits them by description, iterates in one conversation, and reads your schema for business terms.
Describe a chart and AI Copilot builds it. Attach a datasource and it suggests 4 sample queries.
Change the title or colors of the selected widget, or of one named by its title, or color a value with conditional formatting.
"Now group it by region" continues the thread. A session ends after 2 hours of inactivity.
Clear table and column names improve accuracy, and synonyms map "revenue" to a field like "TotalSales".
Good to know: Copilot works from your data's schema, so clear table and column names help. Anything it does not generate is done manually in the Designer.
Examples
AI Copilot prompts are plain-English requests like the three below. For each one, Bold BI generates a SQL query against your connected data.
"Show total sales by month."
Builds the widget from your connected data with the SUM aggregation and picks a fitting chart type.
30-day free trial. No credit card required.
Security & governance
AI Copilot runs inside the Bold BI security model you already use: the permission check happens before the AI response is generated.
For visual generation, the model receives your question, schema metadata and relevant context.
Role-based access control and row-level security are applied on every AI request.
Generated SQL runs only inside the secure Bold BI environment.
SSO: OAuth and SAML sessions carry into AI features in Cloud and On-Premises.
Data sent for visual creation: your question, the datasource schema (table names, column names and data types), and relevant context such as synonyms and conversation history.
Bring Your Own Key (BYOK): on-premises users can use their own OpenAI, Azure OpenAI or Anthropic (Claude) key; Anthropic needs v16.1.90 or later. Cloud Edition uses the Bold AI Service.
AI usage visibility: administrators review AI consumption by user, site and time period.
Audit visibility: on-premises administrators can monitor AI usage and audit activity in the AI-powered Dashboard Usage Insights dashboard; cloud coverage may differ.
AI Copilot like and dislike feedback is stored for 90 days to improve performance.
Limitations
AI Copilot is part of the AI Assistant, so its documented limits apply. Build these manually in the Designer.
FAQ
Yes. AI-generated visuals can contain errors, such as a wrong field, filter or aggregation. Review them before relying on them; every AI visual stays editable.
Yes. AI-generated visuals are standard Bold BI widgets. Change the chart type, filters, colors or fields as you would on any widget.
AI Copilot builds and refines individual visuals in the Dashboard Designer. The in-dashboard AI Agent is the read-only viewer experience and does not edit dashboards. Prompt to Dashboard turns one prompt into a starter dashboard of 6 widgets.
It uses your table and column names as schema context, and synonyms map business terms to field names. Quoting values and using exact column names also help.
COUNT, distinct COUNT (COUNTD), SUM, MIN, MAX and AVG. MEDIAN, VARIANCE and other functions are not supported.
On-premises users can bring their own key for OpenAI, Azure OpenAI or Anthropic (Claude); Anthropic needs v16.1.90 or later. Cloud Edition uses the Bold AI Service. Your own provider bills you directly.
Yes. It runs on Cloud and on-premises, including Docker and Kubernetes. The AI Assistant experience embeds with the embedded SDK with no third-party AI branding; standalone embedding needs version 14.1 or later.
With the Bold AI Service, every AI action uses credits from the license. AI Copilot shows tokens used in the session, today and the remaining balance.
Articles
Guides on AI Copilot, AI credits, embedding the AI Assistant with the SDK, and conversational analytics.
Ready to get started
Try Bold BI AI Copilot free for 30 days: 145 data connectors · Cloud, on-premises, Docker, or Kubernetes deployment.