Case Studies » Leading Financial Institution

Speeding quality assurance

Building an automation framework to accelerate QA in ULAD

As part of its mission to support mortgage markets in the United States, a government-sponsored enterprise (GSE) in the secondary mortgage market supports and manages the Uniform Loan Application Dataset (ULAD).

The ULAD is a recent rewrite of a loan advisor tool used within and without the government to help assess underwriting risks. The ULAD dataset is used by researchers and practitioners to better understand loan trends and to create and manage more persuasive products with the overall goal of greater access to homeownership.

The GSE works within a scaled Agile framework (SAFe) to achieve modern delivery using Agile and DevOps principles. On average, they release 18-24 product increments each year. Quality assurance (QA) for these incremental releases was primarily performed manually, with minimal sets of automated testing for QA. Celerity was asked to build an automation framework to accelerate the QA process for the application with the overall goal of speeding time to market.

Services

  • Agile Development
  • DevOps
  • Process optimization
  • Technology modernization

Services

  • Agile Development
  • DevOps
  • Process optimization
  • Technology modernization

As part of its mission to support mortgage markets in the United States, a government-sponsored enterprise (GSE) in the secondary mortgage market supports and manages the Uniform Loan Application Dataset (ULAD).

The ULAD is a recent rewrite of a loan advisor tool used within and without the government to help assess underwriting risks. The ULAD dataset is used by researchers and practitioners to better understand loan trends and to create and manage more persuasive products with the overall goal of greater access to homeownership.

The GSE works within a scaled Agile framework (SAFe) to achieve modern delivery using Agile and DevOps approaches. On average, they release 18-24 product increments each year. Quality assurance (QA) for these incremental releases was primarily performed manually, with minimal sets of automated testing for QA. Celerity was asked to build an automation framework to accelerate the QA process for the application with the overall goal of speeding time to market.

Using modern delivery methods to speed quality assurance

Celerity started the engagement by developing a modern delivery plan to deliver accelerated QA processes based on industry best practices in DevOps. Deploying a team of test automation specialists and testers in conjunction with business analysts, system analysts, and developers, we broke the problem in to six workstreams.

We first identified the end-to-end business process within the old application and proposed a test automation strategy to validate and verify ULAD’s functionality on three levels—the application programming interface (API) system-to-system (S2S), graphical user interface (GUI) and business process levels—using the advanced toolset within a DevOps/Agile model.

The engagement also included developing and delivering a behavior driven development automation framework using Cucumber and Java Rest Assured libraries for S2S functionality including data validations, loan assignments, messages, feedback certificates, loan functions and protractor with JavaScript for GUI functionality. We also developed standalone and pipeline jobs in Jenkins for continuous integration and delivery.

Reducing quality assurance time with our automation framework

By applying DevOps and modern delivery principles within the client’s SAFe methodology—as well as automating test scripts for end-to-end business processes—Celerity was able to substantially reduce QA time and increase speed to market.

In all, we automated around 700 scenarios at the system-to-system level and around 300 scenarios at the GUI level within the new ULAD application. Compared to a 4-6 week regression cycle using largely manual effort with the old application, Celerity achieved a one-week regression cycle using automated means, significantly reducing time to market.

Before Celerity

ULAD4-6 weeks regression cycle, primarily using manual effort

ULADThree environments to maintain—development, system integration testing and user acceptance testing

ULADNo continuous integration or delivery

ULAD20% automation of UAT test cases

With Celerity

ULAD1 week regression cycle with 95% of the process automated

ULADSingle, continuous integration environment to host and maintain

ULADContinuous integration and delivery enabled through Jenkins

ULAD95% automation of UAT test cases

Leading Financial Institution

Enabling data analysis

Leading National Credit Union

Improving the mortgage underwriting process