Calls recordings go through a number of steps:
1. An agent or customer initiates the call
2. Zendesk Talk creates a new ticket in Zendesk Support
3. The call is completed
4. Zendesk Talk adds a new comment to the ticket it created in Zendesk Support containing a complete recording of the call
5. A trigger notifies Intelligent Voice that a new recording has been added
6. Intelligent Voice retrieves the call recording, transcribes it and scans for potential payment card data.
7. If payment card data is suspected, Intelligent Voice will create a new call recording with the payment card data - all numbers and months - replaced by silence, and a square-bracketed Comment "[Intelligent Voice redacted possible PCI data in this recording]". If your installation fo Intelligent Voice has been set up to display the transcript, the transcript is added to the ticket as a separate new comment, with redacted payment card data replaced by ••••• characters.
8. If no payment card data is suspected, Intelligent Voice will add the original recording and a square-bracketed Comment "[Intelligent Voice scanned this recording for PCI data]". If your installation fo Intelligent Voice has been set up to display the transcript, the transcript is added to the ticket as a separate new comment,
9. Either way, Intelligent Voice will delete the original comment that had been added by Zendesk Talk which contained just the call recording. As such only a single recording of the call, which has been processed by Intelligent Voice, will be left on Zendesk Support ticket.
10. As a belt and braces check, to comply with PCI DSS rules, we recommend you configure Zendesk Talk to delete the original recordings after the minimum available retention period, which is 1 day.