You want the service to be able to verify you're actually gone (not just on vacation, not just changing phones, not incapacitated temporarily). But you don't want it to know what it's guarding or who you are. Those two goals pull in opposite directions. The more the service knows about you, the better it can distinguish "dead" from "busy." The less it knows, the safer your secrets are.
That's a genuinely hard protocol problem. Multi-party verification, tiered escalation, trusted contacts as oracles, ... there are approaches, but nobody's cracked it cleanly yet. The 12-word anonymous account is a solid starting primitive, though.