How do you implement dynamic PII masking across a highly decentralised data mesh without destroying the analytical utility of the data for downstream machine learning workloads?

Likely follow-ups

• →How would deterministic tokenisation need to change if a customer exercises their right to be forgotten and their token must become permanently unresolvable?
• →What happens to your policy engine if a domain team introduces a new data product whose schema doesn't map cleanly onto the existing tagging ontology?
• →How do you detect a data scientist quietly re-identifying tokenised records by joining against an external dataset outside the mesh's control?