Case studyKYC and AML automationFinancial services · Compliance2024

Automating 78% of manual KYC and AML work at Flagstar, from concept to ROI in three months.

Flagstar's BSA/AML and KYC associates spent more than half their time collecting and comparing data rather than assessing risk. We assessed the processes, built the business case, automated the highest-impact work and set up a Center of Excellence the bank now runs itself.

Customer
Flagstar (formerly New York Community Bank)
Size
About $58B in assets, 240+ branches
Division
BSA/AML and KYC
Scope
Assessment, automation and CoE
Outcomes · Reported
78%
of manual work automated in the first phase.
3 months
from concept to ROI.
30+
processes identified for automation over the next 18 months.

Figures are reported results from the first phase of the engagement.

The key finding

Associates weren't short of judgment, they were short of time. Automating the data collection, not the risk decision, gave that time back to fraud assessment.

01 · The challenge

Analysts spent their time gathering data instead of assessing risk.

Every new and existing loan needed borrowers assessed for fraud. Associates pulled information from multiple systems, compared it by hand and triaged alerts before any real risk assessment began.

At the same time, digital fraud was growing more complex, regulators expected more, and the bank needed to onboard and service borrowers faster.

a

Manual collection. More than half of associate time went to gathering and comparing data

b

Rising fraud complexity. Digital fraud made each review harder

c

Regulatory exposure. Any lapse risked fines, legal challenges and reputational damage

d

Slow service. Manual reviews delayed onboarding and servicing for borrowers

02 · What changed

Data arrives assembled, so analysts start at the decision.

Before
Alert raisedData gathered by hand from many systemsCompared and triaged manuallyAnalyst time spent on collection
Risk review starts late
After
Alert raisedData aggregated automaticallyBusiness rules applied, scorecards createdAnalysts focus on true fraud cases
Risk review starts immediately

Digital workers now collect data from internal, external and public sources, apply the business rules, build scorecards and triage cases, so people see only the cases that need judgment.

03 · The approach

We started with the five processes that would prove the value fastest.

Through our Envision assessment we reviewed more than 15 processes and selected five for immediate impact: due diligence searches, paid check verification, internet searches, check deposit verification and the commercial real estate scorecard.

The solution works at two levels: aggregating data from every relevant system, then applying business rules to compare it and produce triaged cases for fraud and exceptions.

04 · Why it holds

The bank built the operating model to keep going.

We helped Flagstar establish its automation program across the seven pillars of a robotic operating model, set up the Center of Excellence and trained the team to be self-sufficient. The first phase paved the way for more than 30 processes across commercial lending operations over the following 18 months.

Where does compliance work pile up?

Show us how alerts and due diligence move through your team today. We will help find the collection work that can come off your analysts first.