Issue #15 | August 4, 2026 | 7 min read
The Brief
When NatWest opened an account for Fowler Oldfield, a Bradford jewellery and gold-trading business, in November 2012, it expected around £15 million a year to move through it. By the time West Yorkshire Police raided the company and shut it down in 2016, close to £365 million had passed through that single account. £264 million of it was cash, some of it delivered to branches in bin bags.
That's the headline number. The part worth sitting with is how it kept happening for nearly four years at a bank with a functioning AML program on paper. The answer wasn't a policy failure or a rogue relationship manager overriding controls. It was a mapping error in NatWest's own automated monitoring system, one that nobody caught, checked, or tested against real account activity until regulators did it for them.
Main Feature
ENFORCEMENT ACTION
The Bank's Monitoring System Read Cash as Cheques. A Jeweller Moved £365 Million Through the Gap.
NatWest didn't miss most of the warning signs on Fowler Oldfield's account. Its own system was built not to see them.
Fowler Oldfield was a jewellery and gold-trading business based in Bradford. NatWest opened the account in November 2012 with an expected annual turnover of roughly £15 million, a number based on the kind of business it was supposed to be. Within about a year, deposit volume started climbing fast. By 2014, Fowler Oldfield was, in NatWest's own internal terms, the single most lucrative customer relationship the bank had in the Bradford area. At its peak, deposits ran as high as £1.8 million in a single day. By the time West Yorkshire Police raided the business in June 2016, roughly £365 million had moved through the account, £264 million of it in cash.
The finding that actually matters here isn't that a jeweller was laundering cash. It's how a bank kept accepting it for almost four years without the account's risk rating ever catching up to what the account was actually doing. NatWest's automated transaction monitoring system had a classification error: it read Fowler Oldfield's cash deposits as cheque deposits. Cheques carry a lower money laundering risk score than cash by design, because they're traceable, bank-to-bank instruments with a paper trail. Every time that misclassification fired, a deposit that should have triggered cash-intensity scrutiny got scored like it came from a chequebook instead. Nobody set out to build a blind spot. Somebody just never tested whether the system was labeling transactions correctly, and it ran that way for years.
£365M — total deposited into Fowler Oldfield's NatWest account, Nov 2012 to Jun 2016
£264M — of that, in cash
£15M — NatWest's expected annual turnover for the account at onboarding
£264.8M — the criminal fine, reduced by a third for NatWest's guilty plea (down from £397.2M)
The system gap wasn't the only failure. Branch staff who physically handled the deposits reported some of it themselves: cash that reportedly carried a musty odor, a volume of Scottish banknotes turning up at English branches with no obvious business reason for it, individual behavior at the counter that read as suspicious even without any compliance training. Those reports went into the bank. Almost nothing came back out the other side. A monitoring system that can't correctly classify a transaction type is a technical failure. A technical failure sitting behind staff who flagged the account anyway, with still no consequence, is a program failure, not a system bug anymore.
Red flags in this case
A commercial account's deposit volume exceeding its stated expected turnover by more than 20 times within roughly two years, with no revisit of the account's risk rating
Cash physically transported and presented in bin bags rather than through any standard commercial cash-handling channel
A concentration of Scottish-issued banknotes deposited at branches with no geographic or business connection to Scotland
An automated transaction-type classification that had never been independently tested against real account activity for the accounts it was scoring
Front-line staff reporting specific, sensory-level suspicion (odor, behavior) that produced no escalation outcome
If your program's confidence in its monitoring output depends on trusting whatever the transaction-type field says a deposit is, this case is the argument for checking that assumption directly, not on a periodic model validation cycle but as a standing question: does the system classify transactions the way you think it does, verified against a reconciled sample of real activity, not the vendor's documentation. Expected-versus-actual turnover reviews on commercial accounts are supposed to be one of the cheaper, more mechanical pieces of ongoing monitoring. This account blew past its number by more than 20 times and never tripped one.
NatWest's guilty plea didn't happen because a compliance officer made a defensible call on a hard judgment case. It happened because a mapping error nobody looked for ran for four years, and £264 million in cash got scored like it came from a chequebook.
Source: UK Financial Conduct Authority, "NatWest fined £264.8million for anti-money laundering failures," December 13, 2021 | FCA, "NatWest Plc pleads guilty in criminal proceedings" | Sentencing remarks, FCA v National Westminster Bank Plc, Southwark Crown Court (Mrs Justice Cockerill), December 13, 2021
Intelligence Briefing
INTELLIGENCE BRIEFING
FinCEN — On July 24, FinCEN issued Alert FIN-2026-Alert004, urging financial institutions to identify, prevent, and report suspicious activity tied to fraud schemes targeting federal student aid programs. Source: FinCEN, FIN-2026-Alert004, "Fraud Schemes Targeting Federal Student Aid," July 24, 2026.
OFAC — On July 15, OFAC designated seven individuals and entities in an international network supporting weapons procurement for Iran's IRGC, following Iran's attacks on commercial vessels in the Strait of Hormuz. The action builds on May and June 2026 designations that targeted the same procurement network. Source: U.S. Department of the Treasury, "Treasury Targets Global Network Procuring Weapons for Iranian Regime," July 15, 2026.
Career Intel
CAREER INTEL
The failure at the center of this case wasn't a policy gap or a bad judgment call under pressure. It was a mapping rule in an automated monitoring system that nobody had tested against real account activity, and it ran undetected for close to four years.
That's a different job than KYC or SAR writing. Model validation and monitoring-rule testing, checking whether the system actually does what its documentation says it does, is a compliance lane where a background in data analysis or QA testing often transfers more directly than a background in banking. Programs that got burned by exactly this kind of gap are why that lane keeps growing.
Tip Line
TIP LINE
Got any idea for something you'd want to see covered here? Doesn't have to fit neatly into a category. If you think it's interesting, we probably will too.
Send it to [email protected].
If someone forwarded this to you, welcome.
The AML Brief goes out every Tuesday. Subscribe for free and get the Top 10 AML Red Flags cheat sheet as a thank-you:
[BUTTON: Subscribe → theamlbrief.com]
Already subscribed? Forward this to one colleague who works in financial crimes. That's how we grow.
The AML Brief | theamlbrief.com
Disclaimer: The AML Brief is an independent financial crimes intelligence publication. All content is sourced from publicly available regulatory documents, enforcement actions, and published research. Nothing published here constitutes legal, compliance, or regulatory advice. The AML Brief is not affiliated with any financial institution, regulator, law firm, or employer. For advice specific to your situation, consult a qualified attorney or compliance professional.