Case study

A production audit trail, reviewed. Ours.

Engagements ask you to trust an implementation you cannot inspect in advance. So the method this site sells ran where honesty is cheapest and proof is hardest to fake: on our own systems. The findings are published, gaps included.

The question

A business case moved from first contact to working files in one day. Afterward, a simple question: can every claim about that case be reconstructed from machine records alone? Not from memory. Not from prose notes. Not by asking the person who was there. If a platform cannot answer that for one case, it cannot answer it for the control environment your officers certify.

The method

The case was tracked end-to-end into a production CRM journal, then the surrounding systems were audited for the same property. The rules were strict on purpose:

  1. Every fact enters through sanctioned machine interfaces only, with an authenticated actor. No direct database writes. Ever.
  2. Every record carries actor and timestamp, so each claim names who did it and when.
  3. Duplicates are prevented, not discovered after the fact.
  4. Where the original lives, the record points to it: message IDs, file paths, verifiable locations.
  5. The review states what is broken as plainly as what works. Findings are not defects to hide. They are the deliverable.

The findings

14/14

case facts journaled in production, with verified sequence numbers and actors

1/6

systems with a complete command and event ledger live in production

12

findings, of which 3 are blocking-class

0

claims in the report that a machine cannot check

The honest result: the ingredients of a complete audit trail existed. Only one system actually produced it. Emails became prose because no machine surface journaled them. A synchronization component had never been deployed, so a documented capability did not exist in reality. Each gap got a finding, each finding got a location, and the remediation list fell out of the review.

What this means for your platform

The readiness review above is the same exercise your platform gets in week one. Controls mapped to the systems that enforce them. Evidence sampled from machine records. Gaps named in writing before anyone certifies anything. The deliverable is the control matrix and the findings list, not a reassurance.

Working papers. The full review and the standing SOX research report are shared on request, source documents included. Client identities removed. No client data appears on this page; the reviewed systems belong to the author.

See the engagements