GENNEXA.aiINTELLIGENT OPERATIONS
All case studies
Case study · bfsi

APAC tier-1 bank

Automated KYC enrichment + analyst copilots cut average review time from 22 min to 6.5 min.

AI SolutionsAnalytics & BIIntelligent Automation
Headline result
3.4× faster AML reviews
faster AML review throughput
3.4×
faster AML review throughput
31%
false-positive rate (from 74%)
38/day
reviews per analyst (from 11)
The challenge

What was getting in the way.

A tier-1 APAC bank with operations in 12 markets was facing a regulatory enforcement deadline: the MAS examination had flagged AML throughput as inadequate. The compliance team of 240 analysts was each completing an average of only 11 customer due diligence reviews per day.

Alert volumes were rising 18% year-on-year from transaction monitoring expansion. False-positive rates hovered at 74% — meaning analysts spent most of their day ruling out non-issues rather than investigating genuine risk. The bank had precisely 6 months to demonstrate measurable improvement to the regulator.

Pain points
A tier-1 APAC bank with operations in 12 markets was facing a regulatory enforcement deadline: the MAS examination had flagged AML throughput as inadequate.
Alert volumes were rising 18% year-on-year from transaction monitoring expansion.
Our solution

What we deployed.

Phase 1

Gennexa deployed an analyst copilot built on a financial-crime LLM fine-tuned on the bank's own typologies, MAS and FATF regulatory guidance, and 5 years of resolved case history. For each alert, the copilot auto-enriched the customer profile from 14 internal and external data sources: corporate registry, PEP/sanctions lists, adverse media, and correspondent bank data.

Phase 2

The copilot generated a draft risk narrative and surfaced the three most relevant precedent cases from the bank's case history — reducing each review to a read-review-decide flow. Analysts were never removed from the decision loop; the tool accelerated them, not replaced them.

Phase 3

A secondary AI model scored escalation likelihood, prioritising the queue so senior investigators received only the highest-risk cases. Every decision generated a full audit trail meeting MAS record-keeping requirements automatically.

Powered byFinancial Crime LLMNexaFlowIDPAnalyst Copilot
Results

What we measured.

3.4×
faster AML review throughput
31%
false-positive rate (from 74%)
38/day
reviews per analyst (from 11)

Average review time dropped from 22 minutes to 6.5 minutes — a 3.4× throughput increase — without adding a single headcount. Daily reviews per analyst rose from 11 to 38.

False-positive escalation rates fell from 74% to 31% as the model learned the bank's risk appetite over the first 60 days. The bank closed the MAS examination gap in 4 months — two months ahead of the regulatory deadline. Senior investigators were able to shift 60% of their time from routine reviews to complex typology development and relationship management.

Get similar results.

Book a 30-minute workflow audit and we'll map the highest-impact opportunity for your operation.

Book a workflow audit