The Generative Answers Logs page lets you explore the inner workings of your Generative Direct Answers (GDA) configuration. It provides detailed logs of the response generation process for each search, so you can debug issues and optimize response accuracy over time.
To access it, navigate to Analysis > Generative Answers Logs in the left sidebar.
Understand Log Statuses
Each log entry shows whether a generative answer was successfully returned for a given search term. Below are the possible statuses:
- Answer Found — The LLM successfully generated and returned a summary to the searcher.
- No Answer Found — The LLM was unable to generate a summary based on the available entities and summarization rules. The process completed, but no usable result was produced.
- Server Error — The summarization process could not be completed within a satisfactory timeframe and was cancelled before returning a result.
Log Detail View
Clicking any log entry opens a detail panel with additional information about that search, including:
- Search Term — The query entered by the end user.
- Searched On — The date and time of the search.
- Status — Whether an answer was found, not found, or errored.
- Search ID — A unique identifier for the search event.
- Total Duration — How long the generation process took, in milliseconds.
- Locale — The language locale of the search.
- Answer — The AI-generated summary that was returned (if one was found).
- Sources — The entity content that was used to generate the answer.
- Referrer URL — The page where the search originated. Use this to investigate where a query came from and troubleshoot response quality in context.
The detail panel also includes a Generation Process tab, which shows a step-by-step breakdown of how the answer was generated.
Exporting Logs
Click Export in the top-right corner to download your logs as a .csv file. The export includes all standard log fields, plus:
- Referrer URL
- Sources
- A direct link to the log detail in the platform UI