Every external review engagement has a deadline. And in the weeks before that deadline, something predictable happens: the reviewer's behaviour changes. Files that would have been queried get rejected outright. Rejection rationales shift from specific to generic. The workload that was supposed to be shared between reviewer and institution gets quietly redistributed — almost entirely to the institution's side.
Nobody documents this formally. Nobody talks about it in the exit meeting. But anyone who has worked through the end of an external SOW review cycle knows exactly what it looks like on the ground.
The Rejection Pattern
As an engagement approaches its close, a subtle triage logic takes over. Files that are straightforward get processed. Files that require significant effort — complex structures, layered wealth sources, incomplete primary documentation — get rejected. The stated reason is almost always the same: insufficient primary documents to complete the review.
That rationale is not always wrong. There are genuinely files where the documentation base is too thin for a meaningful review. But the threshold for what counts as "insufficient" tends to lower as the deadline approaches. A file that would have generated a detailed query letter in month two of an engagement generates a blanket rejection in the final two weeks.
The rejection is not a finding. It is a deferral dressed in the language of a finding. The file's problems don't go away — they get handed back to the institution to solve under time pressure that didn't exist before.
The practical effect is that the hardest files — the ones that needed the most rigorous external scrutiny — are the ones that end up unreviewed. The ones that sail through are the easy ones. The external review has, in effect, quality-selected for simplicity at exactly the moment when complexity needed the most attention.
The Standard Drift
It's not just rejections. The quality of the review work itself changes as the engagement winds down. Queries become shorter. Rationales become thinner. Commentary that would have been specific — "the savings rate assumption of 60% is not supported given the client's stated lifestyle expenditure and property ownership" — becomes generic: "SOW narrative requires strengthening."
When you push back on a generic query, the response is predictable: the RM didn't provide enough information for the reviewer to be more specific. That may sometimes be true. But it is also a convenient explanation that places the accountability entirely on the institution side while the reviewer's output gets thinner by the week.
The Hidden Cost — Who Actually Pays
Here's what the engagement metrics don't capture. When a file gets rejected in the final two weeks, it doesn't disappear from the institution's obligations. It comes back — to the governance team, to the RM, to whoever is responsible for ensuring the SOW review cycle is completed. And it comes back at exactly the worst time.
The governance team is not just handling rejected files. They are simultaneously running new-to-bank onboarding, periodic reviews, exit trigger reviews, and all the other BAU obligations that don't pause because an external engagement is winding down. Nobody formally assigns the overtime. Nobody formally requests the weekend work. It happens because the alternative is a material gap in the review cycle that will show up in the next audit.
The external reviewer's engagement metrics show files processed and findings issued. They do not show files rejected under questionable rationale, commentary that thinned as the deadline approached, or the downstream remediation burden absorbed by the institution's internal teams. The engagement looks complete on paper. The people who worked nights and weekends to make it look complete are invisible in the record.
What Good Engagement Management Looks Like
This is a governance problem that is solvable — not by working harder at the end, but by managing the engagement differently throughout. Three things make the difference:
- Track rejection rationale quality from the start — not just rejection rates. If rejection letters start becoming generic three months before the engagement ends, that's the signal to escalate, not to remediate. Generic rejections should be challenged formally and early, not accepted as a standard.
- Front-load the complex files — the files most likely to generate difficulty should go into the review queue first, when the reviewer's attention and appetite for detailed work are highest. Leaving complex cases for the back end of an engagement is a structural error that guarantees end-of-cycle pressure.
- Document the reviewer's output quality as a formal record — if commentary is thinning, note it. If rejection rationale is generic, note what specific questions were not asked. This creates a factual basis for engagement debrief discussions rather than a vague sense that things weren't handled well.
The Structural Truth Nobody States Plainly
External reviewers are not neutral actors with no incentive to manage their own workload. They have engagement economics — hours budgeted, team bandwidth, contractual deliverables defined in terms of files processed rather than quality of output. As the deadline approaches, the rational move for a reviewer under time pressure is to process the easy files and return the hard ones. The institution's governance team has no equivalent exit. They carry every file that comes back.
This is not a criticism of any individual reviewer or firm. It is a structural feature of how external review engagements are designed and priced. The fix is not to work harder in the final two weeks. The fix is to recognise the incentive structure early and manage the engagement accordingly — which means the institution, not the reviewer, has to be the one tracking quality throughout the cycle.
The people who end up working weekends at the end of a review cycle are not doing so because of their own failures. They are absorbing the accumulated consequences of documentation gaps, RM shortcomings, and reviewer triage decisions that were made across a months-long engagement. The right response is not to normalise that pattern — it's to understand it well enough to break it before the next cycle starts.