Example: Discover Decision Points
Discover decision points for a loan underwriting application
Main Description

The interviews with the underwriter, subject matter expert, gives the rule development team the following information:

  • The loan application contains information about the applicant (his name, his employment status, etc...) in a special form called the 1003 form, also known as the Uniform Residential Loan Application form.
  • The 1003 also contains information about the mortgage type that the applicant has selected (interest rate, term, purpose, etc…) and about the property (property type, address, etc…).
  • The underwriter starts by looking at the 1003 to perform some basic verification on the application, such as checking that the age of the primary applicant is at least 18, or that the primary applicant is a permanent US resident. If any of these conditions is not met, the application is rejected.
  • If the application passes the initial validation, the underwriter then gathers the credit data for the primary borrower and the co-borrower if there is one. This consists in retrieving the applicants detailed credit information, such as details on their different lines of credits (e.g. credit cards, car loan, possible bankruptcies) and their previous mortgage history (possible foreclosure and mortgage lates) from several external services called credit bureaus. From the combined information, the underwriter calculates the value of different parameters such as the credit score and the risk grade.
  • The company offers several loan programs relying on different risk criteria: ScoreTrust is the most popular program and relies on a combination of credit score and mortgage lates grading.
  • The credit score considered by the company is the middle of the 3 score provided by the credit agencies or the lower of the score if there is less than 2 reporting agencies available.
  • Then, the underwriter reviews the eligibility of the borrowers for the selected loan program. The eligibility step is assessing:
    • The risk is taking in lending money to the applicants using in particular the credit score and the risk grade determined from the credit information.
    • The validity of the loan with respect to certain state laws or to company specific guidelines.
  • The assessment will result in either accepting the application, referring it for further (manual) review, or rejecting it:
    • If the risk is too big, the loan application will be referred for further review (e.g. the applicants may be proposed another type of loan).
    • If the loan does not comply with the state guidelines or the company own guidelines, the application will be rejected.
    • If the risk is acceptable and the request is conform to the eligibility guidelines, the loan request is accepted and a corresponding LTV will be determined from LTV guides sheets for each program.
  • During the eligibility verification step, a set of stipulations and conditions messages (for example: "Cash-on-hand must be verified back to the source.") are also collected and will be reported on the underwritten loan offer.
  • The next step is the pricing activity. The purpose of this step is to determine the final rate of the loan, using the rate sheet for the selected loan program and applying different adjustments based on property and applicant characteristics. It is important to record the breakdown of the pricing (i.e. the value of base rate, the different adjustments, etc…) for audit purpose.
  • The pricing for a loan is determined using the appropriate loan program rate sheets. These rate sheets are tables which give the base rate for a given credit score, risk grade and LTV (Loan To Value). The rate sheets also define the different adjustments.
  • Once the loan is priced, the complete response (including the loan amount, the rate, stipulations and conditions) goes to the Loan Origination System for closing.

From these information we can draw a high level process map which will help us to extract some decision points. We do not need to detail the tasks for the purpose of this sample.

From a first study we can see that there are a lot of decisions and expertise in the data validation and eligibility steps, that for sure are important decision points to work on. So the discovery roadmap will be articulate to organize workshops on those two steps. In fact the others steps like "Calculate Risk grade and credit score", "select loan program", and "pricing" are also candidate for rules discovery, as risk management and pricing includes a lot of exceptional policies a company may want to have the flexibility to change often to adapt to competition and market changes. This loan application processing sample is following a simple business process, we will design using a process map. So we will use a business process oriented discovery approach to model the loan application processing.

The decision points table may look like:

Decision Point Name


Source for Rule Discovery

CurrentStateof Automation

Rule Owner -


Data validation

Validate the information entered around the borrower, the amount requested, the property

1003 forms

Legal document per state and federal

Some rules are implemented in the current GUI used to enter the data In the system


Calculate Risk grade and credit score

calculates the value of different parameters such as the credit score and the risk grade

Score cards

Excel files of the underwriter



Select a loan program

Loan program can be defined as a set of conditions a borrower and the property may need to respect.

Program definition and stipulation




Review the eligibility of the borrower for the selected loan program

Underwriting policies




Price the loan

Pricing list

Score cards

There is an existing application