Fundamentals of Credit Decision Engine 

A Credit Decision Engine is a specific type of Rules or Decision Engine that memorializes a bank or a financial institution’s directive in terms of granting end users an application for a bank account or a line of credit.

In the past 30 years, banks have had the exclusive need for decision engines. But banks didn’t develop these decision engines with features and needs for other institutions to use. Most of the time they hard coded these rules and decisions with archaic programming languages that’s no longer supported. The rules which they implemented are also exclusively for the use of the banks and aren’t meant to be flexible.

Fast forward to today’s explosive FinTech era, the need for decision engines and specifically decision engines used for credit and underwriting decisions are prevalent. However, the technical aspects of the decision engine still haven’t caught up. 

Fundamentals of Credit Decision Engine - Connectivity

Connectivity is probably the most important aspect of a well thought out decision engine architecture. Some of the decision engines available today claim to be a turnkey solution, but with further examination, there isn’t anything turnkey about it. Users still have to spend god awful amounts of time to integrate third parties and augment the decision engine’s infrastructure and miss credential deadlines of their product launch.

A well thought out decision engine especially for credit risk management must come with all of the critical data vendors already connected. From an identity risk perspective, TransUnion's TruValidate, Sentlink and Socure are amongst the top performers in the identity verification space. A modern decision engine must have some of these third party identity data vendors already connected and the user just has to drop in their credentials and make it work.

Connectivity does not stop here, there are other types of critical connectivities such as banking core system, credit card management system and other types of enterprise platforms. When a decision is made the decision engine’s credit decision must be routed to the next step to either inform the user and their end client of an approval, decline or needing for additional information.

Fundamentals of Credit Decision Engine - End User Centricity 

Some of the commercially available decision engines are fundamentally insulated and do not communicate with any other systems let alone the end-user that’s interacting with the decision engine through an online application or online journey.

This failure of end user centricity is causing confusion with the user and end-user of the credit decision engine and will cause a mountain of compliance issues. 

Another fundamental if not critical connectivity of a credit decision engine is the ability to generate actions and trigger an outcome. An example of actions are approval, decline, review and request for additional documentation or information. A decision engine should use these actions to trigger different outcomes, one of the outcomes could be a message displayed on the online experience and another outcome could be a text message or email letting the end user know what’s going on. Lastly the decision engine should communicate with the end-user’s user portal if there are any further actions to be worked on.

Some of the turnkey solutions out there lacks these fundamental connectivities to properly communicate credit decisions made within the decision engine and therefore make the entire customer journey very confusing and the users such as banks and credit unions will have no clue what’s being decided and what was communicated to the end user as well as themselves on the next steps.

Fundamentals of Credit Decision Engine - Variables and Rule Chaining

Now that we understand the importance of connectivity, we also need to understand the mindset of the user of credit decision engines. Some of these users are risk managers and most of them have an excellent idea of how they want to set up their rules and how they want these rules to be executed. 

However, most of the risk managers and risk officers must rely on engineering staff to help them to implement rules and worst integrate variables coming from these third party data providers mentioned in the previous section of the article. Often, some of the turnkey solutions do not go deep enough with their implementation and leave out nuances that a risk management professional needs to do their work.

A great credit decision engine must have third party integration completed, it must also have a fundamental understanding of all of the content that’s supplied by these third party data providers. In the instance of TransUnion TruValidate, there are hundreds of variables and many scores tuned to identify identity fraud for a risk manager and risk officer to use.

A variable library is a must for risk management staff to select from and write rules with. For example if a risk management professional wants to write a rule to decline anyone with a fraud risk score of less than 500 (lower the score, the higher chance of fraud). He or she should drag and drop the fraud risk score into the authoring tool and simply move over a math operator to the effect of (Decline of Fraud Score <= 500).

If that fraud score is not made available but the integration is somehow finished with some of the turnkey solutions, it’s essentially useless and requires additional engineering and a few months of back and forth to get just one variable prioritized into a work queue. And precious product launch time is again delayed.

Fundamentals of Credit Decision Engine - Visualization

Another fundamental characteristic of a modern credit decision engine is a graphical user interface. The days of writing lines of code and using somewhat low code solutions such as Drools are over.

A decent decision engine should have a graphical user interface for risk management professionals to author rules as well as visualize all the rules of various nodes, end points and transition points to other rules.

Some of the users of LendAPI’s visual decision engine have created a vast web of rules and become fairly complicated for their business needs and a visual rules builder helps the rules engineers to visualize the complexity and interaction of all of the rules they’ve built.

Check out LendAPI’s Credit Decision Engine

When we constructed LendAPI’s credit decision engine, we had all of these fundamentals in mind and we made all of these connectivities, variables available for a risk management professional to jump in quickly, build their rules, link various outcomes and test their entire ruleset. Visit us and build your own rules free for 30 days at www.lendapi.com and click on the Sign Up button on the upper right hand corner.