A personal loan product running in 2 weeks, down from months

Many fintechs and new lenders entering a new market wanted to offer personal loans. Most did not have an experienced credit team in house. Building a product from scratch meant understanding two things at once: the country's lending regulations, and what borrowers in that market actually want. That took months.

The team built a ready-made personal loan product to close that gap. A lender starts from a working product with a default risk appetite already built in, and tunes it to its own.

Stage	What it does	Status
1Study what works	Compares existing personal loan products	Done
2Set the guardrails	Works within the country's lending regulations	Done
3Design the engine	Defines the variables, logic and scorecards	Done
4Write the standard document	Records it in one standard, AI-readable format	Done
5Configure and deliver	Sets it up in the engine, ready to deploy	Ready

The starting point was a cross-analysis of several personal loan products already running for clients. It compared which input variables each one used, how their rules were built, and where they differed. The common patterns became the base of a best-practice product.

Consumer lending in the market is governed by the central bank's regulations. The product was designed under that guidance from the start, so a new lender does not have to work it out alone.

Every part of the credit logic was designed for the product. That covers which input variables to collect, the formulas and logic built on them, affordability through the debt burden ratio (DBR, the share of income that goes to debt repayments), the decision trees and matrices, and the scorecards with their maths. A scorecard is a points system that turns an applicant's details into a risk score. A default risk appetite is built in, so the product makes sensible decisions on day one. The lender then adjusts cut-offs and limits to its own appetite.

The product is written up in one standard business requirements document (BRD). It is standard in structure and in its visual layout: the same sections, tables and decision tree diagrams in the same order. That makes it easy for a credit team to review, and easy for AI tools to read and update later.

The product is not only on paper. It is already configured in the decisioning engine. A lender can run it on the provider's cloud, or host it on its own infrastructure. Delivery follows a set plan: discovery calls, configuration in the lender's engine, a few rounds of iteration together, then go-live. The product is running in about two weeks and live within a month.

• About 2 weeks to a running personal loan product, instead of months from scratch.
• Live within a month, including discovery, configuration and iteration with the lender.
• 60 to 70% ready on day one. The lender tunes the rest to its own risk appetite.
• No need for a full in-house credit team to design the first product.