Skip to content
All case studies

Finance · 2024

Replatforming a national lender for real-time decisions

We replaced a 20-year-old decisioning core with a cloud-native platform that approves loans in seconds, not days.

A Fortune 500 consumer lenderFinancial services · 6,000 employees

2.4sMedian decision latency, from 3 days
99.98%Decisioning availability during cutover
3.1xExperiment throughput for pricing models
0Origination days lost to migration

Client

A Fortune 500 consumer lenderFinancial services · 6,000 employees

Technology stack

AWSKubernetesKafkaPostgreSQLGoFeature flagsOpenTelemetry

ROI

312% over 24 months, payback in 7 months.

Client overview

A national consumer lender was losing share to faster digital challengers. Its decisioning core, two decades of accumulated logic on a mainframe, could not meet modern latency, audit, or experimentation requirements.

Business challenge

Migrate mission-critical decisioning off a mainframe without halting originations, while improving latency from days to seconds, keeping every decision auditable, and introducing the experimentation the business needed to compete.

Approach

We ran a six-week architecture discovery to define target domains and a strangler-fig migration path, then delivered the platform in increments: a new decisioning service behind the existing interface, a feature-flagged rollout, parallel-run validation, and a controlled cutover with full audit lineage.

Business impact

The lender regained competitive parity on speed, opened a continuous experimentation program on pricing and risk, and cut per-loan infrastructure cost by 41%. The audit lineage built during migration became a durable advantage in subsequent regulatory reviews.

Let's talk about the decision that decides the next decade.

If the problem matters enough to warrant experienced leaders, it matters enough to start the conversation.