Traditionally, UCI performs the credit analysis process through the company’s own risk analysts, who study the processes and make a decision. Although they are professionals with extensive experience and knowledge of the company’s risk policies, the only manual analysis generates some limitations or consequences, such as

Delay in response: they invest a lot of time in review to approve or reject each credit application. Sometimes, it would take days to review a single case.

High competitiveness: unable to offer binding responses in real time, potential customers who make an application often end up signing with the competition.

Presence of subjectivity: the final decision was taken only by a team of people specialised in this area, but with feelings, prejudices; who, despite following some policies, contribute with subjectivity and this can mean that, in some cases, decisions were taken that are not totally neutral or objective. In addition, there have been occasions when real estate agents, who are their main clients, have asked to review transactions on the basis of approval of similar transactions, or have demanded an explanation of the reasons for rejection to justify it to the end client.

Lack of agility to adapt to change: changes in risk policies for mortgage approval are constantly being made based on client behavior or simply because of regulatory changes. Specialized agents need time to learn these new rules and be able to apply them in analysis.

Mortgage default: many of the approved mortgages have not been paid. Apparently, the mortgages were granted to people who did not meet all the requirements.

Faced with this situation, UCI wanted to create a dynamic system capable of evaluating applications in real time to pre-qualify its potential customers and thus reduce the number of defaults.

Habber Tec developed for UCI an artificial intelligence solution that analyses mortgage applications in real time, improving the level of services provided to customers and the company’s results.

The technologies used were IBM Watson Machine Learning, to run the predictive models, combined with IBM ODM (Operational Decision Manager) to manage risk policies through rules, IBM Watson Studio with cognitive services that improve the user experience and capture new data to improve the machine learning models and IBM Cloud Native Development Subscription for IBM Cloud.

It is 100% SaaS based, with a control panel that allows the tracking and monitoring of the KPIs defined for the platform. This solution is easy and with a very low initial investment, economical to maintain and update, as well as a great scalability.

Improvement in the quality of service provided to the client because the solution offers applicants a real-time response and an explanation of what motivated the decision on their mortgage application, so that even if the operation has been rejected, the client can take into account these criteria for a new demand for housing or request for financing.

It increases business capability by automating 80% of credit decisions, considerably increasing productivity and the ability to manage operations without increasing equipment.

Accurate decision making, as decisions are made by state-of-the-art artificial intelligence algorithms that allow reliable predictions, minimising the information that must be requested from the client and providing decisions that can be explained on a case-by-case basis, thus reducing fraud and subjectivity.

Policy compliance, as it incorporates IBM’s cloud-based automation and decision management technology, which allows risk analysts to create and update new rules when risk policies change, so that they can be applied seamlessly across different business processes and applications.

Better use of the time of specialized agents, allowing them more time to define policies and analyze complex cases or exceptions, leaving the technology to manage most operations and all the processing of repetitive tasks.

Decrease in the percentage of non-payment due to a consistent credit approval decision.