Most email verifiers do a regex check, a DNS lookup, and call it a day. That works on roughly two-thirds of the addresses they see and fails on the rest — specifically on the Google Workspace and Microsoft 365 mailboxes that accept every probe (because anti- enumeration accepts every RCPT TO and lets the rejection happen later, post-send). A surface-level verifier reads those as Valid · 95% confidence. Then the real send goes out and the bounce hits the sender's domain reputation.

KillBounce was built specifically to get those right. The probe is three layers — syntax, DNS/MX, and a live in-session SMTP check — with explicit catch-all detection and a category for "we can't tell without sending, here's the risk score." Nothing is labelled valid unless the evidence is there.