How we built Oraculi to help lenders make informed decision

At the core of any decent loan management software is something that helps a lender decide whether a loan should be given. Decisioning is, therefore, a challenge every lender must solve. If you get it wrong, your loan business will fail. 

In the early stages, some of our lenders used foreign decision engines like Credolab, FICO and others. This proved expensive for these lenders because these services were being charged in dollars, the Nigerian economy with its ever-depreciating naira, and the not-that-great NPL turnover of these decision engines. 

As Lendsqr is always in the business of solving problems, we decided it was time to build our decision engine, one that’s cost-effective for our lenders, especially those in the Nigerian market. Firstly, we deliberated on what essentials are expected of a proper decision engine:

• It must be accurate
• It must be fast
• It must be configurable
• It must be cost-effective (AKA cheap!)

So, we built Oraculi — the Latin word for “The Oracle.” Of course, we all know that the oracle sees the future. Now, you might wonder if Oraculi also sees the future. Well, it can glimpse some parts of it, decently better than the foreign ones. But does it see everything 100%? Unfortunately, at this time, it doesn’t. 

Inner-workings and pros of Oraculi

1. Speed: For Oraculi in Lendsqr, we maintained the speed because some of the data is resident to us, like the Karma blacklist. So, for bad loans, we return and decline at the earliest opportunity instead of relying on slow APIs to make a decision.
2. Modular: We made Oraculi modular because we found that lenders don’t need similar risk control parameters or decision complexities for every type of loan. This made it easy for lenders to decide what module to call. Some fancy loans could have about nine modules, while a simple back-office loan could have just one.
3. Cost-effective: By layering the free data services internal to Lendsqr, before sending it to external data providers and credit bureaus, we made Oraculi affordable and cost-saving for all lenders. 
4. Swift configurability: Setting up and configuring a decision model used to take a considerable amount of time. But now, it can be done in minutes.

Modules on Oraculi

Decision modules are system-level components for implementing risk settings for a loan product. The following are the available options within Oraculi.

Karma

This module taps into one of the largest private blacklist databases of bad actors and chronic defaulters. It protects billions of loans for lenders each month. Karma can be configured to check emails, phones, BVN, IP addresses, etc. This is available free of charge to Lendsqr lenders and is available over APIs for non-Lendsqr lenders.

Ecosystem

Ecosystem is the module that checks the entire Lendsqr data ecosystem to see if a borrower meets specific criteria during decisioning. It uses the BVN as the primary key for looking up this data. It checks the borrower’s activity with other lenders on the Lendsqr data ecosystem. For example, it can determine if a potential borrower is owing other Lendsqr lenders. This is similar to credit bureau data but on steroids. Read more about the extensive Ecosystem data dictionary. It is free for Lendsqr lenders and is available over APIs for non-Lendsqr lenders.

Whitelist

The whitelist feature allows a Lendsqr lender to create a specific loan product for a user or a group of users using either the user(s)’s BVN, phone number or email. This loan product will only be available to users linked to the product. The whitelist feature can be used in different cases.

For example, you could have an arrangement with a specific company to give their employees loans at an interest rate that is not open to everyone on your app. Here, bulk whitelisting will quickly create this product and link it to the specific users of this company. The whitelist feature could also be used to run promotions. If you have a set of returning good borrowers that you wish to appreciate, you could do a bulk whitelisting, enabling these users to access a new product only they can. Finally, you could whitelist a single user you may have personally vetted using the single-entry option.

Statement

Statement data includes information such as income, expenses, assets, debts, and a history of past payments and balances. With this information, a lender can determine if a borrower has a stable source of income and if they have sufficient resources to cover repayments. By using statement data, lenders can make informed decisions about lending money and minimize the risk of default. Lendsqr-powered lending apps come with different statement data providers, such as Mono, myBankStatement, and others, which can be easily configured.

Loci

This module allows a lender to write smart rules against high-speed velocity data of customers. For example, check instantly the number of times a borrower has made specific changes to their information, such as loan requests, net income, etc. Read more about the extensive Loci data dictionary with over 80 data points. Loci is free for Lendsqr lenders but is available over APIs for non-Lendsqr lenders.

Scoring

This is the traditional weighted and risk-adjusted scoring method beloved by traditional risk managers, where a lender can apply numeric weights and scores against customer data. Lenders can use available data points within the loan request and other modules, such as Ecosystem or Loci, to make decisions. For example, certain jobs could have higher scores. Customers whose summation of different scores from various parameters are higher than the module’s threshold proceed to the next module. Scoring is free for Lendsqr lenders but is available over APIs for non-Lendsqr lenders.

Periculum

Periculum — an Oraculi custom module provides Lendsqr lenders with advanced ML algorithms to analyze transaction data effectively. With access to classification, random forest, and neural networks, lenders can instantly scrutinize various transaction details, including timestamps, transaction types, device IDs, and IP addresses. This thorough analysis helps reduce risk, enhance accuracy, and proactively identify potential fraud attempts. Periculum can promptly flag suspicious activities that may have otherwise gone unnoticed. Periculum ensures every detail, whether bank or card transactions, is carefully examined, enabling our lenders to stay one step ahead of fraudsters. 

Credit bureau

This module allows a lender to get online real-time data for anyone from any of Nigeria’s credit bureaus. A lender can configure a single or multiple credit bureaus to be searched simultaneously. We currently support the three credit bureaus in Nigeria (including TransUnion): CRC, CreditRegistry, and FirstCentral. This is available for a fee and is charged per use.

Let the Oracle guide your decisioning

As we look ahead, Oraculi promises a transformative shift in lenders’ decision-making. No longer confined to pre-built modules, lenders will soon have the power to custom-create, define, and configure decision modules tailored to their specific loan business needs. 

This customization capability heralds a new age of precision in lending decisions, empowering lenders to lend confidently and shrewdly. Want to lend smarter and save yourself from tons of bad decisions? 

Reach out to us at [email protected], and let’s get you started.