SARs the bank failed to identify and file under its old thresholds, on nearly $4 billion of activity, found in its own one-time lookback. The new order adds an independent SAR lookback that reaches back to January 2022, including alert groups that scored below the investigation line and were never worked.
FinCEN release, $1.2B penaltyAxle Lookback
Find what your triage closed by mistake.
Rose, our transaction monitoring agent, re-works alerts your team already closed: auto-closed, dismissed, and below threshold. You get a report on what the first decision missed, before an examiner goes looking.
Request a Triage Test →
2 weeks. 500 to 1,000 closed alerts. No integration.
Closed alerts by risk scoreIllustrative
Closed, Rose agrees
Closed, Rose would escalate
Why now
Regulators are reopening closed alerts.
Three of this year's biggest AML actions turn on the same question: what did the monitoring program close, or never look at, that it should have reported? A lookback is how you answer it. Better to run one on your terms than under an order.
Of suspected trade-based money laundering over the past decade that the bank did not timely identify, evaluate and report. The consent order requires an independent SAR look-back to find previously unreported activity and check the quality of past filings.
OCC release, $350M penaltyIn administrative monetary penalties, a record. FINTRAC issued 35 notices of violation, up 52% on the prior year, including the largest penalty in its history.
FINTRAC annual reportHow it works
Three steps. Nothing to integrate.
1
Send closed alerts
Export a CSV of closed alerts from your monitoring system, or give us read access to a date range. Include the alert, the customer profile and the transactions behind it. A light security review is all it takes.
2
Rose works each one fresh
Rose investigates every alert as a new case, using your SOP and risk appetite. She does not see the original decision, and nothing is written back to your system.
3
You get an examiner-ready report
Closes Rose would escalate, by rule and score band. Estimated SAR yield above and below your threshold. Every case with Rose's reasoning. Tuning suggestions for the rules that need them.
Three ways to run it
Start with a sample. Scale to the full window.
Triage Test
A two-week test on a sample of your auto-closed and below-threshold alerts. Ends with a report for your CCO.
- Scope
- 500 to 1,000 closed alerts
- Timeline
- 2 weeks from data in hand
- For
- Any team running a TM system
No cost for qualified teamsOr a small fixed fee, credited toward a contract.
Full Lookback
Every closed alert in a 12 to 24 month window, worked as a real case and documented for your examiner or consultant.
- Scope
- All closed alerts in the window
- Timeline
- Set at scoping, by volume
- For
- Teams under exam pressure or remediating findings
Per case, volume-tieredQuoted after a scoping call.
Rule Test
Run a shadow copy of a new rule or a lower threshold and tag what it fires. Rose works those alerts on the side and reports how many would escalate and how many would clear.
- Scope
- One rule or threshold change at a time
- Output
- Escalate vs clear rate and the L2 workload it adds
- For
- Teams tuning rules or lowering thresholds
Per rule testedOr bundled with Rose.
Sample report
What lands on your CCO's desk.
Built so you can hand it to an examiner or your independent tester as is. The figures below are made up to show the format. They are not from a customer.
Axle Lookback · Triage Test
Closed-alert review: Example Bank
Window: 1 Jan to 30 Jun
Population: 41,200 closed alerts
Sample: 800 (stratified by rule)
Prepared by: Rose, reviewed by Axle
Population: 41,200 closed alerts
Sample: 800 (stratified by rule)
Prepared by: Rose, reviewed by Axle
Example figures only. Your report uses your alerts, your rules and your thresholds.
800
Closed alerts re-worked
27
Closes Rose would escalate (3.4%)
9
Of those, likely SAR-worthy
2
Rules driving 70% of misses
Escalation-worthy closes by rule
| Rule | Sampled | Rose would escalate | Rate |
|---|---|---|---|
| Rapid movement of funds, 7 days | 180 | 11 | 6.1% |
| Structuring below reporting limit | 140 | 8 | 5.7% |
| High-risk jurisdiction, inbound | 160 | 4 | 2.5% |
| Dormant account reactivation | 120 | 3 | 2.5% |
| Crypto on-ramp velocity | 200 | 1 | 0.5% |
By score band
| Score band | Sampled | Escalate |
|---|---|---|
| 40 to 49 (just below line) | 260 | 16 |
| 20 to 39 | 310 | 9 |
| 0 to 19 (auto-closed) | 230 | 2 |
Estimated SAR yield
| Population | Alerts to SAR |
|---|---|
| Above threshold (your data) | 2.8% |
| 40 to 49 band (Rose) | 2.3% |
| Below 40 (Rose) | 0.4% |
Case list with Rose's reasoning (2 of 27)
ALT-118203 · Score 46 · Closed 14 MarEscalate
Twelve inbound transfers from nine unrelated senders over six days, moved out within hours to one exchange. Pattern matches pass-through activity. Original close cited "consistent with profile"; the profile lists payroll income only.
ALT-120877 · Score 31 · Closed 2 AprEscalate
Cash deposits of $9,400, $9,200 and $9,600 at three branches in four days, followed by a wire to a newly added beneficiary. Structuring indicators were not addressed in the original disposition.
Tuning suggestions
- Lower the review line on rapid movement of funds from 50 to 42. Expected added volume: about 310 alerts a month.
- Add a counterparty-count condition to the structuring rule so multi-branch cash is scored as one pattern.
- Crypto on-ramp velocity clears 99.5% of the time below the line. Leave the threshold where it is.
Next step: turn Rose on for live alerts, so these closes get a second look before they close.Request a Triage Test
What you get, and what we touch
A read-only review. Your system stays as it is.
- The CCO reportFindings by rule and score band, estimated SAR yield, and what to change.
- Every case, documentedRose's full investigation and reasoning for each alert, ready for your QA team to check.
- A readout with our teamWe walk your team through the findings and the cases that matter most.
- No integrationA CSV export or read access to a date range. No connector, no change to your monitoring setup.
- Nothing written backYour dispositions stay untouched. What you do with the findings is your call.
- Data deleted afterWe delete your data once the report is delivered. SOC 2 Type II and ISO 27001 certified.
Proof
We publish our numbers.
Nobody required it. We asked FairPlay, an independent third party, to evaluate our screening agent and published what it found. Rose runs the same way: your SOP, a full audit trail on every decision, and results you can check.
1,048
Cases scored by FairPlay
97.5%
Overall accuracy
99.2%
Precision
97.7%
Specificity
FAQ
Questions compliance teams ask.
Which monitoring systems does this work with?
Any system that can export closed alerts with the customer and transaction data behind them. That covers the common case managers and in-house tools. If you can pull a CSV, we can run it.
Does Rose change any of our dispositions?
No. Rose works a copy of the data and nothing is written back. The report tells you which closes she would escalate and why. Your team decides what to reopen.
Can this count as the independent lookback in a consent order?
Orders usually require an independent consultant approved by the regulator. In those cases we work alongside the firm you engage, or through them, so their team gets Rose's case work and reasoning to review. For everything short of an order, the Triage Test is yours to run now.
How does Rose know our risk appetite?
We configure Rose from your SOP and escalation policy before the run, and you sign off on that setup. A sample of her decisions is checked against your analysts' calls before the report goes out.
What do we need to give you?
The closed alerts in scope, the customer profile for each, and the transactions in the alert window. A 30-minute scoping call settles the sample and the format.
What happens after the report?
You can turn Rose on for live alerts, so the next borderline close gets a second look before it closes, or run a Rule Test before your next threshold change. Neither is required.
Sources: FinCEN, Bank of America consent order (Oct 13, 2026), sections on the 2024 lookback and the SAR Lookback undertaking. OCC, American Express National Bank consent order (Oct 8, 2026), Article IX. FINTRAC, 2025–26 annual results. FairPlay figures are for Baley, Axle's screening agent, on a standard configuration.
See what your queue closed.
Request a Triage Test: enter your work email, then book a 30-minute scoping call. We will send a data checklist before the call.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
