← All work
Digital lending & credit risk · MENA

Blink by Capital Bank

Scaling a neobank's digital lending fast, without letting credit risk run away with it.

The challenge

Blink, the first neobank in Jordan, set out to scale digital lending quickly. The hard part is that speed and credit risk usually pull in opposite directions: the faster you acquire borrowers, the more exposure you take on. There was already a decisioning flow in place, but it leaned heavily on what customers declared about themselves rather than on hard, verified data, which is exactly where risk hides. My job was to grow the book aggressively while keeping risk firmly under control.

The credit decisioning engine: verified data and AI to a risk score in seconds
The credit decisioning engine: verified data and AI to a risk score in seconds
What I did

As Product & Credit Risk Lead, I redesigned the credit decisioning engine end to end and put verified data and AI at the centre of it. Blink was the first product in Jordan to approve a credit card application entirely online, in seconds. A customer downloads the app, opens an account and applies for the card in a single flow, and the moment they submit, the engine goes to work. It connects directly to the government identity agency to verify the national ID, including its biometric check, and to pull authoritative data such as age, then to the national credit bureau to retrieve five years of credit history, and combines both with the information the customer provided in the app. AI does the heavy lifting on the analysis: it reads across all of these signals together, how a customer's age, credit history, occupation and even the city they live in interact, and learns which combinations actually predict repayment and which predict trouble. That analysis produces a single risk score for each applicant, which maps directly to the credit limit they qualify for. I designed the scoring logic, the API-driven data flows and the onboarding and activation journeys so approvals were fast, automated and grounded in real data, working hand in hand with risk, compliance, engineering and commercial teams.

After launch I built AI-driven reporting and analysis on top of the live portfolio, watching how each segment actually performed and where fraud and scam patterns were trying to creep in. That feedback loop let the bank continuously tweak the model, tightening it against emerging fraud and abuse without slowing down genuine customers, so the engine got sharper and more bulletproof the longer it ran. Off the back of the Jordan launch, the proposition was set up to expand into Iraq as the next market.

Applying for a credit card in the app, approved in seconds
Applying for a credit card in the app, approved in seconds
The result

More than 20,000 users acquired within six weeks of launch, and over $12M in high-risk lending exposure removed through data-driven credit controls and risk scoring. Rapid growth and credit discipline, at the same time.

Let's talk product.

Always happy to talk product, AI, and interesting opportunities.