Troubleshoot Google Cloud Dialogflow CX and ES integrations

AI-generated summary

Genesys Cloud provides integrations with Google Cloud Dialogflow CX and ES to enable conversational AI capabilities. If you experience issues configuring these integrations, first troubleshoot based on your agreement: For Google Cloud Dialogflow agreements, open a support case with Google. For Genesys Cloud agreements, open a case with Genesys Cloud Customer Care. Common errors include exceeding quotas, interactions taking the failure path in Architect due to no input timeouts, and recognition issues where calls don't go to expected queues. Review training phrases, intents, and no input event handling. If troubleshooting does not resolve the issue, open a case with the appropriate support team.

If you experience difficulties configuring your Google Cloud Dialogflow ES or CX integration, review these troubleshooting tips to determine whether you should open a support case with Google, or a Genesys Cloud Customer Care case.

I have an agreement with Google for my Google Cloud Dialogflow service

Open a Google Cloud Dialogflow support case. For more information on Google support, review these articles:

I have an agreement with Genesys for my Google Cloud Dialogflow service

Open a Genesys Cloud Customer Care case in My Support Portal. For more information, see About Genesys Cloud Customer Care.

Common Google Cloud Dialogflow CX or ES errors

You may encounter Google Cloud Dialogflow service errors if a high number of interactions take the failure path, or if interactions encounter recognition issues.

Note: If the following actions persist after you troubleshoot the application, then open a case with Genesys Cloud Customer Care.

Interaction takes the Failure path in Architect

Possible reason	Description	Action
Quota issues	The following metrics exceeded the configured number: Dialogflow Essentials Edition audio query operations per minute of service “Dialogflow.googleapis.com” limit Dialogflow All other requests per minute of service “Dialogflow.googleapis.com” limit All requests per minute of service “texttospeech.googleapis.com” limit	Review the Quotas section of your Google Cloud account. If you exceed the quota, open a case with Google Cloud Dialogflow support.
Google Cloud Dialogflow ES only: Maximum number of No Input timeouts occurred	The connection closed, possibly because the Dialogflow application never handled any of the following events: GENESYS_INITIAL_NO_INPUT GENESYS_NO_INPUT GENESYS_FINAL_NO_INPUT If the intent that handles GENESYS_FINAL_NO_INPUT is not set to “End of conversation,” then after the event triggers, the call takes the failure path. If the intent that handles the GENESYS_FINAL_NO_INPUT is set to “End of conversation, then the call flows the flow that matches the intent in Architect, and not the failure path.	Review the conversation recording and make sure that the customer is not uttering the correct intent, or is using DTMF. Make sure that you assign events for your Google Cloud Dialogflow bot to use when a caller does not respond to a prompt. For more information, see Configure Google Cloud Dialogflow intent behavior when the caller does not answer. If the configured intent behavior does not correct the problem, open a case with Google Cloud Dialogflow support.
The Call Dialogflow Bot action or the Call Dialogflow CX Bot action do not return a result.	Google services might be down.	Check the Google Cloud status dashboard.

Possible reason

Description

Action

Quota issues

The following metrics exceeded the configured number:

Dialogflow Essentials Edition audio query operations per minute of service “Dialogflow.googleapis.com” limit
Dialogflow All other requests per minute of service “Dialogflow.googleapis.com” limit
All requests per minute of service “texttospeech.googleapis.com” limit

Review the Quotas section of your Google Cloud account.

If you exceed the quota, open a case with Google Cloud Dialogflow support.

Google Cloud Dialogflow ES only:
Maximum number of No Input timeouts occurred

The connection closed, possibly because the Dialogflow application never handled any of the following events:
- GENESYS_INITIAL_NO_INPUT
- GENESYS_NO_INPUT
- GENESYS_FINAL_NO_INPUT
If the intent that handles GENESYS_FINAL_NO_INPUT is not set to “End of conversation,” then after the event triggers, the call takes the failure path.
If the intent that handles the GENESYS_FINAL_NO_INPUT is set to “End of conversation, then the call flows the flow that matches the intent in Architect, and not the failure path.

Review the conversation recording and make sure that the customer is not uttering the correct intent, or is using DTMF.

Make sure that you assign events for your Google Cloud Dialogflow bot to use when a caller does not respond to a prompt. For more information, see Configure Google Cloud Dialogflow intent behavior when the caller does not answer.

If the configured intent behavior does not correct the problem, open a case with Google Cloud Dialogflow support.

The Call Dialogflow Bot action or the Call Dialogflow CX Bot action do not return a result.

Google services might be down.

Check the Google Cloud status dashboard.

Interaction encounters recognition issues

Error	Description	Action
Calls do not go to the expected queue.	The Google Cloud Dialogflow bot may encounter a tuning or recognition issue and the customer input did not match an intent. For example, your intent might be to perform an action when the customer says, “Yes,” but the customer actually says, “Yep” or “Yeah.”	Review the Training section of your Dialogflow bot and assess the quality of your trained intents. For more information, see Google Cloud Dialogflow’s Training phrases page. If you adjust the intent and the problem persists, open a case with Google Cloud Dialogflow support.

Error

Description

Action

Calls do not go to the expected queue.

The Google Cloud Dialogflow bot may encounter a tuning or recognition issue and the customer input did not match an intent. For example, your intent might be to perform an action when the customer says, “Yes,” but the customer actually says, “Yep” or “Yeah.”

Review the Training section of your Dialogflow bot and assess the quality of your trained intents. For more information, see Google Cloud Dialogflow’s Training phrases page.

If you adjust the intent and the problem persists, open a case with Google Cloud Dialogflow support.

[NEXT] Was this article helpful?

Get user feedback about articles.