Voice AI debugging information
Several environmental factors might prevent Pega Voice AI from working as expected. Use the strategies described in this section to debug issues with Voice AI in your Pega Customer Service instance.
Pega Customer Service Implementation Guide
The two key types of issue related to Voice AI that might require debugging are:
- Data flow issues, which prevent the system from transcribing the captured audio and sending it to the Pega server.
- Agent desktop issues, in which the agent desktop does not receive the expected suggestions despite the availability of transcripts.
Before you begin debugging:
- Verify that the VerifyVoiceAIDataRetrieval activity returns a 202 or 204 result. If the activity does not return this result, the configuration is not complete, and the debugging steps in this section might not work.
- Enable desktop debugging. Only errors are sent to browser console logs by default. Set the EnableAIDesktopLogging when rule to return true, to enable additional logging to the browser console, including logging of VAD events, utterances, suggested cases and articles, and detected entities.
- Enable debugging to the
PegaRules.logfile. Use the EnableVoiceAIPartialUtteranceLogging and EnableVoiceAIFinalUtteranceLogging when rules to add additional logging to the
PegaRules.logfile of the RealTime node. If you set these when rules to return true, then partial or final utterances are printed to the logs. Enabling either of these rules also prints a count of the transcripts in each run of the data flow.
- Ensure that you have access to the RealTime node to watch the tail of the
PegaRules.logfile, and access to the developer tools on the browser, to watch incoming events to the desktop.
Previous topic Voice AI configuration testing Next topic Data flow issues