PROBLEM: False positives create noise, and admins won't adopt AI they can't verify.
APPROACH: Put the emphasis on what AI removed from classification rather than what it kept, because that is where the value is.
Hello! I’m Supriya
Product Designer | B2B SaaS | 5+ years of experience
Currently designing experiences at Microsoft shipping security, compliance and AI features from 0 to 1.
PROBLEM: False positives create noise, and admins won't adopt AI they can't verify.
APPROACH: Put the emphasis on what AI removed from classification rather than what it kept, because that is where the value is.
PROBLEM: Admins needed to protect older sensitive data, but pay-as-you-go scanning made the cost uncertain.
APPROACH: Made the cost estimate mandatory, and offered "Download report" instead of "Start classification", because admins needed budget approval before they could commit.
PROBLEM: Admins re-entered and updated the same long lists rule by rule, across policies.
APPROACH: One central list library shared across all DLP workloads, with a guided migration from Endpoint DLP.