We just enabled data masking; POST June 30, 2026 improvements. We had no choice in enabling it, as the system does not allow editing of Requestor notes, which is what we did in WHD to manually redact PII. Years ago, we did the same thing at PayPal in email responses. Point is, the current level of data masking is incredibly over-powered. For example, our Department names, Team names, and Agent names are all being masked in outbound emails, which looks pretty unprofessional. But there's no way to avoid it based on our legal requirements.
My only suggestion is a "Safe" list for specific types of info (names) or text strings (department names). Or perhaps a sensitivity slider that would control how strict the AI is??? I don't know enough about how it works to provide truly helpful tips on how to modify the agent.
