Turn Existing Dashboards into Interactive AI Analytics Experiences

The Challenge

Business dashboards make important metrics visible, but they often provide only predefined views. When users need to investigate an unexpected change or ask a new question, they may need SQL skills, analyst support, or a separate report.

This can limit how easily teams explore their own data:

• Dependence on technical skills: Answering questions outside predefined views can require SQL knowledge or analyst support.
• Limited exploration: Fixed charts and metrics do not anticipate every follow-up question a business user may have.
• Slower decisions: Teams may wait for custom queries or reports before they can investigate performance changes.
• Underused data: Useful information may remain difficult to access even when the underlying data is already available.

The Approach

Add an agentic conversational analytics layer to an existing dashboard. Rather than relying on one predefined query, the agent can inspect the available data, run focused queries to investigate different parts of a question, combine the results, and produce a complete answer in context. GitHub data demonstrates the pattern.

• Let users ask questions in natural language, without needing to write SQL for every investigation.
• Allow the agent to discover relevant tables, schemas, and field values before choosing how to investigate the question.
• Break complex questions into focused data checks, run multiple queries where useful, and synthesize their results into one coherent answer.
• Use query results and execution feedback to refine the investigation or recover from query errors when needed.
• Apply appropriate data-access permissions and query safeguards for the connected data source.
• Preserve familiar dashboard charts and KPIs while adding conversational exploration for questions those visuals do not answer on their own.