Methodology · module 01, screening

How we measured −70%.

Same dossiers. Manual route versus Compl.AI, logged in the officers' own KPI journals, reconciled line by line against production data.

What was counted, what was not, where it breaks, and the protocol to run the same test on your cases.

−70%
average across the officers
manual versus Compl.AI
875 → 278 min
pooled, on the same dossiers
≈43 h
officer time returned
across the pilot

scope · 875 → 278 is the paired dossiers. the 43 hours is the whole pilot. two scopes, not one number stated twice.

Same dossier, both routes.

The officers kept paired journals. The same case done the manual way, and done with Compl.AI.

A dossier entered the set on the officer's own sign-off. The bank chose, not us.

Real cases off a live production queue. Not a benchmark set.

Every journal line was reconciled against production data. A logged time had to match a screening that actually ran.

We publish no dossier count and no per-officer figure. Eight officers in one bank are identifiable.

The measurement, stated plainly.

Unit
officer hands-on minutes on one dossier
Platform run time
counted as officer time, not excluded. The clock kept running while the machine worked
Comparison
the same dossier, manual route versus Compl.AI
Source
the officers' own KPI journals, kept during live work
Reconciliation
every line matched against production data, not against memory
Population
8 compliance officers · 4.5 weeks · 300+ screenings · live production at a leading Swiss bank
Statistic
the average of the officers' own figures, manual versus Compl.AI, not a pooled total
Machine work
the platform read 63,019 articles in full across 37 languages. no officer read them
Not counted
our engineering. 210 improvements implemented during the pilot, on our clock, not theirs
Not claimed
accuracy. a number from our own test set is not evidence to your model-risk function
Not published
dossier counts, per-officer figures, the bank's name

Why an average, not a pooled total.

Each officer is one observation. Averaging officers stops the heaviest files deciding the headline.

One case in the pilot carried 1,993 articles. Pool everything and cases like that one write the number.

Pooled, the same dossiers ran 875 minutes to 278. Slightly less than −70%. We say so rather than pick.

Measure quick cases in their intended mode.

Bank-approved deterministic rules can fully automate small, low-risk checks.

Keep them in the pilot and measure straight-through volume, exceptions and officer time.

What we will not claim.

No accuracy number. A figure from our own test set is not evidence to your model-risk function.

So we hand over the method instead, and let your officers produce the number.

Conclusions were compared too, not just clocks.

Speed alone proves nothing. Each pair was also checked for the conclusion it reached.

Where the two routes diverged, the case was root-caused. Every one of them, during the pilot, not afterwards in a report.

Each was fixed and put into production straight away. That loop, not the time saving, is what the officers said they valued.

210 improvements implemented · 4.5 weeks · reported issue to production in minutes to hours

We publish no count of pairs and no agreement fraction. Eight officers in one bank are identifiable and that is their business, not our marketing.

Run it on your own cases.

Ours ran 4.5 weeks in live production. Keep week one in: the learning curve is part of the cost.

01
Pick off the live queue

Your real cases, your mix. Include the quick low-risk checks, so the spread shows instead of hiding.

02
Pair every dossier

Same case, both routes, same officer. Alternate which route goes first, so knowing the answer cancels out.

03
Define the clock before you start

When it starts, when it stops, whether waiting on a run counts. Decide once, in writing, for both routes.

04
Log the minutes and the conclusion

Both, on every dossier. A time saving without the decision attached to it proves nothing at all.

05
Reconcile against system records

Weekly, line by line. Journals are memory. Reconciliation is what makes them evidence.

06
Root-cause every divergence

Where the two routes disagree, in writing, before anyone quotes a result from the exercise.

What would prove us wrong.

A test you cannot fail is not a test. Any of these sinks the claim on your book.

Officer minutes do not drop on EDD, complex cases and adverse media.
that is where we say the value sits · claim fails
Something your manual route caught is missing from the surfaced set.
recall · disqualifying, not a trade against time
The gain sits in a few large files and your median case is unchanged.
spread · then the average oversold it

One more, whatever the clock says. If a decision cannot be replayed months later, the time saved is worthless.

The trace is what makes that checkable: what the file keeps.

Bring your own cases.

We will run this protocol with you and report what it says, including if it says less than −70%.

Questions on the method: contact@complai.ch.

Book a scoping call