Value Packs / Banking / Deposit-Flight & Fraud Guardian / Brief 01

Deposit-Flight & Fraud Guardian · Brief 01

Suspicious FD breaks · Branch 07

Affluent-uninsured · behaviour → money → book

DataGenie Brief · autonomous
Branch 07 · affluent slice · cross-dataset lineage · in your cloud

The one line you configure

Which branches show FD breaks and outbound transfers running out of band, and is it rate-driven flight or theft? Trace it across the three books.

In a nutshell

Branch 07 carries all of it, and it reads as suspicious money movement, not pricing or service. Premature-broken FD jumped to $306.8K vs $31.4K expected, with 5 ticketless breaks where 0 were expected; CASA net flow fell to −$169.9K as outflow hit $476.7K, with $296.0K going to first-time payees and $149.2K over instant rails. Rate-shopping stayed low and complaints stayed at zero, ruling out yield-chase and service attrition. It is one pocket, not the franchise: affluent-uninsured, high-flight-risk deposits.

At a glance

FD Premature Broken$306.8Kvs ~$31K normalBranch 07 · affluent slice
FD Break Rate75.9%vs 26.0% expby count · +5 ticketless (0 exp)
Net CASA Flow−$169.9Kvs +$20.6K exp= the $476.7K out less the FD credit
Outflow Value$476.7Kvs ~$165K normalhigh-risk
First-Payee Outflow$296.0Kvs ~$6K normalnever-seen payees
Rate-shop / complaintslow / 0discriminator→ theft, not price/service

Connected signals — what else moved (across all three books)

Instant-Rail Outflow$149.2Kvs ~$26K normalfast rails
Contact-Change Count2vs ~0.4 normalATO tell
Micro-Test Intensity11.10 expectedpre-fraud probe
FD Ticketless Breaks50 expectedcontrol tell
Complaint Rate0flatnot service
Rate-Shoppinglowflatnot price

Engagement (contact-changes, micro-tests) and CASA (instant-rail) moved with the FD-break — the cross-book corroboration; complaints and rate-shopping stayed flat, the discriminators that name it theft.

FD Premature Broken — actual vs expected band

FD Premature Broken ($K) · actual vs expected band~10× normal
0K100K200K300K400K~10× normaltoday
Actual$306.8K
Expected$31.4K
Ticketless breaks5 (0 exp)

Behaviour → funds → outflow → destination

Behaviour (leads)Source of fundsOutflowDestination
Contact-changes ×2 · micro-testsengagement leadsFD broken $306.8KBalance draw $169.9KCASA$476.7Kall outflowFirst-time payees $296.0Knever-seen · $290.1K above normalEstablished payees $180.7Kpaid before · only $21.5K above normal

Behaviour leads; FD liquidation plus a balance draw fund a $476.7K outflow, and $149.2K of it moved on instant rails. The two destinations look comparable in width but they are not: established payees are only $21.5K above their normal week, while first-time payees are $290.1K above. Nearly all of the abnormal money went to people we have never paid.

Where the extra $311.6K came from — expected $165.1K → actual $476.7K

$0K$119.18K$238.35K$357.53K$476.7K$165.1KExpectedoutflow+290.1First-timepayees+21.5Establishedpayees$476.7KActualoutflow

These bars are the increase over expected, not the totals — same $476.7K as the flow above, split a different way. Of the $311.6K over baseline, first-time payees are $290.1K of it: $296.0K actual against $5.9K expected, a 50× jump. Established payees, the ones we have paid before, added just $21.5K on top of their usual $159.2K. Outflow finished at 2.9× its baseline.

What the connected view adds

Each dataset alone looked ordinary; only stitching engagement + CASA + FD on the same Branch 07 / affluent slice surfaces the break-to-extraction chain on day one — the window a periodic review misses.

So what

Read together, the three books tell one story: behaviour signals lead (contact-changes, micro-tests), the FD book is liquidated ($306.8K broken, ~10× normal), and the cash leaves as a $476.7K outflow — $296.0K to first-time payees, $149.2K on instant rails. Rate-shopping low and complaints at zero rule out price and service. Caught on day one at Branch 07, in the affluent-uninsured, high-flight-risk pocket.

The points that matter

Lead with the amount, not the rate

$306.8K broken vs $31.4K expected, with 5 ticketless breaks where 0 were expected.

$306.8KFD broken

The money left to strangers, fast

$296.0K to first-time payees (vs ~$6K normal) and $149.2K on instant rails — money being taken out, not a deposit maturing.

$296.0Kfirst-payee

The flat signals name it theft

Rate-shopping low and complaints at zero rule out yield-chase and service; only the break-rate and fraud tells moved.

0complaints

One pocket, not the franchise

Concentrated in affluent-uninsured, high-flight-risk deposits at Branch 07; the region only mirrored the branch-led spike.

Branch 07affluent · high-risk

Recommendation

Route Branch 07's first-payee and instant-rail destinations to fraud investigations and hold the linked beneficiaries; make the behaviour → money → book chain a standing daily cross-branch watch. DataGenie surfaces and explains — the freeze / return decision stays with your fraud desk.

Questions it already answers

Have these same payees received money from any other branch?

How much more sits in the affluent-uninsured pocket at Branch 07?