The Recommendations page displays AI-generated recommendations for the analysis of a recording. The AI generates Key Findings as part of its recommendation response.

The page displays the summary of Key Findings with the recommended action items, each with a priority, an action, and a rationale. Progress recommends that you review the recommendations before you apply them to your application business logic.
Note: The tool generates recommendations automatically when an analysis is created, if the analysis status is included in recommendations.autoRun.healthStatuses in the analysis.config file. An empty status list disables automatic requests, but you can still request recommendations from this page.
If recommendations are not available, the page displays a message and a Request AI Recommendations button. Click the button to request recommendations for the analysis summary of the recording. Each recommendation request sends the generated analysis summary, with the application-managed prompts, to the pre-configured AI endpoint. It does not send the recording file or raw recording data.
Note: Recommendations require an AI service, but the Analysis Summary does not. In the initial release, as of OpenEdge 13.1, the default service is Progress Agentic RAG using the OpenAI Chat Completions service pattern. Other LLM providers have not been tested at the time of this release. The configured Progress Agentic RAG endpoint is provided in the llm.config file. To use AI recommendations, you only need to supply a valid apiKey. If the key is missing or incorrect, or if an error occurs during a request, the tool displays an error message through its standard notification system.