Reporting Assistant: Automated Follow-Up Questions and Feedback
The Reporting Assistant now generates a list of automated follow-up questions after producing a report summary, helping users decide what to explore next based on the data in their report. A thumbs up/thumbs down feedback mechanism has also been added at multiple interaction points to help Cornerstone continuously improve the application.
What's New
- Automatically generate follow-up questions after report insights
- Click a suggested question to fetch deeper data instantly
- Provide thumbs up/thumbs down feedback on report lists, insights, and suggestions
- Specify feedback detail such as helpful, accurate, clear, confusing, incorrect, missing info, or too complex
Details
Description
The Reporting Assistant has been enhanced to support a more guided, conversational analytics experience. After a user selects and runs a report, the system acknowledges that the report is running and generates an insight. With this release, the LLM then produces a list of three follow-up questions the user might want to explore, derived directly from the data contained in the report.
Selecting any suggested question prompts the system to fetch and return the relevant information. For example, if a report contains position data, the assistant may suggest exploring which positions have the highest number of training versions, then return a specific answer such as the position with the most training occurrences. A second enhancement introduces a feedback mechanism, allowing users to rate suggestions and insights with a thumbs up or thumbs down and to specify whether the output was helpful, accurate, clear, or whether it was confusing, incorrect, missing info, or too complex.
Business Need
Clients do not always know what questions to ask when looking at their data. Without guidance, valuable insights buried in a report can go unexplored, limiting the value users derive from their reporting efforts.
By having the LLM suggest relevant follow-up questions, the Reporting Assistant lowers the barrier to deeper analysis and helps users uncover insights they might not have thought to pursue. The feedback mechanism ensures Cornerstone can continuously gauge the quality of report lists, insights, and suggestions, driving ongoing improvement of the application over time.
Key Features
- The system automatically generates a list of three follow-up questions after report insights are produced, based on the data contained within the report.
- Users can click any suggested question to have the system instantly fetch and return the corresponding data.
- A thumbs up/thumbs down feedback control is presented at multiple points: when a list of reports is provided, when a report insight is delivered, and when an individual suggestion is answered.
- Users can specify the nature of their feedback, indicating whether output was helpful, accurate, or clear, or whether it was confusing, incorrect, missing information, or too complex.
Deployment & Considerations
These enhancements are part of the Reporting Assistant closed beta rolling out in April, building on the March release that stopped after generating the initial insight. The top three closed beta capabilities are the conversational interface for requesting reports, automatic insight generation for reports, and the ability to converse with users.
As a closed beta feature, access is limited to participating customers rather than generally available. Administrators should review the considerations and exclusions provided with the release before enabling the feature. Standard Reporting Assistant prerequisites, including appropriate reporting permissions and access to the underlying report data, are expected to apply.