Name:
Ricardo Orjuela
Juan Esteban De La Calle
Diego Alexander Maca
Country: Colombia
Organization: Davivienda
Davivienda is a regional bank leader that operates in six countries (Colombia, Panama, Costa Rica, El Salvador, Honduras, and the United States). With over 50 years of experience in the Colombian market, Davivienda offers a wide range of financial services to individuals, SMEs, and corporate customers. The bank currently serves over 20 million customers through a network of more than 650 branches and 2.500 ATMs.
Awards Categories:
Business Challenge:
As one of the main banks in Colombia, we constantly face the challenge of cluster and prioritize our customers in an effective and efficient way to optimize our resources and efforts in key areas of the business, such as call centers, risk management, fraud, and marketing.
We realized that, even though many areas in the bank had their own segmentation and prioritization systems, these systems often were inconsistent and were based on variables selected subjectively instead of statistical analysis and AI techniques.
For example, the credit risk management area selects its best clients based on credit score. Meanwhile, the credit placement area selects discount on interest rate based on the amounts the client has within the bank.
These ad hoc approaches not only led to inconsistent results between different areas of the bank but also were inherently limited, given that they did not capture the complete gamma of information available to evaluate our customers. This created friction between our processes, given that the lack of a prioritization and segmentation made our efforts and resources difficult to optimize.
To overcome these challenges, we decided to begin a project to develop a comprehensive customer score. Our expectation with this project was to create a centralized punctuation system that would use advanced analytics and AI techniques to evaluate our customers in a comprehensive and precise manner. We thought that this focus would allow us to segment and prioritize our customers in a more effective way, helping us to improve resource allocation and improve the efficiency of the bank.
Business Solution:
To solve our challenge of segmentation and prioritizing our customers, we appealed to Dataiku, given its solid data manipulation capabilities, its interoperability with our existing infrastructure, and its robust support for advanced AI techniques and analysis.
Our team, composed of seven members, including data scientists, business analysts, and personnel from our CDO, worked together to implement the solution. We used Dataiku’s capabilities to connect with Cloudera to obtain data from our 20 million customers (banking and users from our digital wallet).
Our focus consisted of developing a personalized optimization technique based on the maximization of a multiple ROC AUC, defined from a set of binary variables dependent on the multiple endogenous variables we wanted to optimize simultaneously. This technique was designed especially for our need to obtain a comprehensive and precise score of our clients which would focus on the sorting between clients.
The process in Dataiku was end-to-end. We start with the extraction and preprocessing of data, move through the development and testing phase of the model, and finish with the implementation and monitoring of the model in production. All of this was done inside the same platform, which let us maintain full traceability and made collaboration easier among team members.
One of the standout benefits of using Dataiku was speed. It let us process big volumes of data efficiently and quickly. Also, the platform provided an environment that made checking the results easier, which increased our trust in the insights generated by the model.
In summary, Dataiku was crucial in our ability to develop and implement an advanced AI solution that satisfied our needs for customer segmentation and prioritization and helped us surpass the challenges we faced.
Day-to-day Change:
The implementation of the "Comprehensive Customer Score" project has deeply changed our day to day in the organization. Before this project, each department had its own method of prioritizing customers, which could cause inconsistencies and a fragmented view of our customers. With our new scoring system, we have a joined and coherent approach to prioritizing our customers in the organization.
The score has become our first choice when we have to prioritize customers, no matter the channel or the project. This integral score gives us a full and joined view of each customer, allowing us to make informed and precise decisions in all the areas of our business.
The clearness and integrality of this new way of prioritizing customers have positively impacted our business processes and decision-making across the organization. We can now make sure that we are making decisions based on a complete evaluation of each customer, allowing us to optimize our resources and efforts.
In summary, the implementation of this project has given our organization a new way of working and making decisions, aligning all departments with a single and consistent customer prioritization. This has brought greater consistency, efficiency, and effectiveness to our daily operations.
Business Area: Analytics
Use Case Stage: In Production
Value Generated:
The implementation of the comprehensive customer score has generated significant value for our organization. In financial terms, investment return has been big, making around 2.6 million US dollars, with very low cost because we used existing resources.
Beyond financial ROI, the value generated by this project extends to many areas of our organization. With a unified scoring system based on AI, we have avoided the need for each department to make its own score, leading to better operational efficiency and saving of resources.
In addition, this unification has improved consistency in decision-making across different departments, helping to improve strategic alignment and ensure all decisions about client prioritization are based on the same information and methodology. This change has led to better resource allocation and more efficient decision-making, contributing big value to our organization.
In short, implementing this project has created quantitative value in terms of ROI and provided big qualitative benefits in terms of efficiency, consistency, and informed decision-making across our organization.
Value Brought by Dataiku:
Dataiku has brought great value to our organization in many ways.
Efficiency and speed: With Dataiku, our team has been able to work better and faster. The platform has let us handle big amounts of data quickly and well, improving our ability to process and analyze information. This has made us faster in generating insights and making decisions based on data.
Integration and technology efficiency: Dataiku fits well with our existing infrastructure (Cloudera). Working together has improved our technology, letting us use our existing resources and reducing the need for different solutions or doing the same work twice.
Risk management and transparency: Through its ability to give complete tracking and enable teams to work together, Dataiku has made us better at managing risks and maintaining transparency. This has let us develop and keep a complete view of our data operations, making governance and risk control better.
Skills development and networking: The Dataiku Academy and Dataiku Community are valuable resources offering opportunities for skills development and connecting with other data and AI professionals.
In short, Dataiku has provided important value by improving team and technology efficiency, managing risk and transparency, and offering potential opportunities for skills development and networking.
Value Type:
Value Range: Millions of $
