Least privilege is easy to agree with and hard to keep. The usual failure is not a missing policy but a slow one: when the approval path takes a day and the workaround takes a minute, associates choose the minute, and the policy becomes a document.
The model that works is just-in-time access with an expiry. Standing access is reduced to what a role needs every day. Everything else is requested when needed, approved in the tool the approver already has open, granted for a bounded period, and revoked automatically. The Evidence Engine handles the grant and the revocation; the human only decides.
Speed is the product. A request in Privr puts three things in front of the approver on one screen: what is being asked for, what the requester has used before, and what the policy says about that combination. Most approvals take under a minute; low-risk requests that fit a pre-approved pattern are granted with no human at all, with the pattern recorded as the approver.
Expiry changes the economics. Because every elevated grant ends on its own, nobody has to schedule cleanup, and the quarterly review becomes a list of standing grants only. At a mid-sized law firm the standing grants fell from 1,900 to 400 in two quarters, not because anyone was stricter but because nothing accumulated.
The exceptions are where the signal is. When one person requests the same elevated access every week, the policy is wrong, not the person. The Evidence Engine flags the repeats and proposes a standing grant for review, so policy drifts toward the work rather than away from it.
Evidence is a by-product. Each grant carries its request, approver, policy version and expiry, so the GDPR question about who had access to case files on a given day is answered by a query over the Audit Vault, not by an email thread.
Just-in-time elevation is part of every Privr plan. Request a walkthrough on your own approval flow.