Enhancing the User Experience Exponentially with a Robust Control Tower System for Alerts and Predictions
One of the largest financial institutions in South America, serving over 17 million clients, reached out to Igerencia, a Nimble Gravity Company, for assistance in addressing significant issues in their digital channels and transaction processes, which were directly impacting customer satisfaction.
To resolve current problems and mitigate future issues, Igerencia, a Nimble Gravity Company, designed and implemented an analytical framework for monitoring through a series of smart control panels. These panels perform predictive analysis using AI models and generate monitoring alerts, facilitating an organized and efficient tracking of financial transactions and digital channels.
Results:
- 360° monitoring from a single indicator.
- 13.8 thousand clients monitored per minute.
- 60% reduction in response time.
- Improved platform availability.
Technologies and Libraries Used:
Apache Spark |
Impala-helper |
PowerBI |
Numpy |
Python |
Pandas |
Cython |
Pyspark |
Fbprophet |
Pystan |
Holidays |
Xgboost |
Full Case Study
Title: Enhancing the User Experience Exponentially with a Real-Time Control Tower System for Alerts and Predictions
Client: Bancolombia
Sector: Banking and Finance
Project: Control Tower to monitor transactional and digital channels in real-time.
Bancolombia, the largest financial services institution in Colombia, needed an experienced data analytics partner to enhance the performance of their digital channels.
As one of the largest banks in South America, Bancolombia provides banking and financial services to over 17 million clients. In Colombia, the bank focuses on improving access to financial products and services to promote economic development. Through its branch network and digital platforms, Bancolombia strives to offer a secure, efficient, and reliable banking experience. Additionally, the bank supports many community development initiatives and promotes financial inclusion in rural and underserved communities.
Despite being one of the country’s largest banks, Bancolombia previously lacked the technical capacity to tackle significant issues in its digital channels and transactions, which directly affected customer satisfaction. Due to the high volume of complaints, the bank was concerned about the risk this posed to its reputation and profitability, a potential disaster in a highly competitive market.
Bancolombia approached Igerencia, a Nimble Gravity Company, to design and build a comprehensive and secure monitoring solution using data analytics, capable of predicting and preventing critical situations. They also needed robust dashboards with precise customer segmentations for interactive and visual reporting.
The Control Tower of Alerts Uses AI Technology to Deliver Predictive Analysis and Transaction Alerts
Igerencia, a Nimble Gravity Company’s vast experience in Microsoft environments, including Azure and Databricks, along with over two decades of experience implementing business intelligence and data analytics systems, made them the ideal partner for a project of this scale, which included complex requirements and security risks. The team’s initial goals were to ensure data integrity, promote security, and enable informed decision-making, building trust with the financial institution and strengthening its relationship with customers.
Igerencia, a Nimble Gravity Company, designed and implemented an analytical framework for monitoring through a series of smart control panels. These panels perform predictive analysis using AI models and generate monitoring alerts, facilitating organized and efficient tracking of financial transactions and digital channels.
The system combines four specific modules:
Analytical Module: Used to predict banking transaction outcomes within a time range. The analytical module uses historical data and mathematical models for predictions. It also handles data entry and extraction from HDFS tables.
Probability Module: Used to calculate an indicator that determines whether a banking transaction’s completion code should have appeared based on its probability during analysis against the KPI set by the business. It receives HDFS tables as input sources and stores the results in the same format.
Alerts Module: The alerts module receives results from the first two modules, compares these figures with business data, and classifies them into three levels based on deviation: normal (green), warning (yellow), and attention needed (red). For data extraction, processing, and storage, the first three components use Python and Impala as the SQL query engine. Due to the large data volumes, the analytical and probability modules run only during low platform usage, avoiding additional problems for end users.
Visualization Module: Uses Power BI to graphically display data from the alerts module. Dashboards contain valuable information on successful versus failed transactions, failure reasons, and more. It provides users with powerful report filters, making it useful for most stakeholders and technical teams.
The Control Tower Provides 360° Monitoring, Unprecedented Channel Availability, Improved Response Times, and a Significant Increase in Customer Satisfaction
360° Monitoring from a Single Indicator: The comprehensive monitoring solution offers a complete and detailed view of operations through a central indicator. With this tool, the client can comprehensively monitor all relevant areas and processes, allowing them to quickly and accurately identify any situation requiring attention.
13.8 Thousand Clients Monitored Per Minute: With the new platform, the bank can monitor 13,800 clients per minute simultaneously, giving them unprecedented access to prevent events and provide personalized, timely service at all times.
60% Reduction in Response Time: The complete monitoring system and optimized processes have led to a significant reduction in response times, notably improving customer satisfaction.
Improved Platform Availability: By anticipating critical situations and improving response times, service availability is higher than ever, ensuring that all channels are ready to meet the needs of every client.
Testimony
“The implementation of the comprehensive monitoring system and optimized processes has led to a significant 60% reduction in response times, greatly improving customer satisfaction.”
“The dashboards contain valuable information on successful versus failed transactions, failure reasons, and more. They provide users with powerful report filters, making them useful for most stakeholders and technical teams.”
“Igerencia, a Nimble Gravity Company’s vast experience in Microsoft environments, including Azure and Databricks, along with over two decades of experience implementing business intelligence and data analytics systems, made them the ideal partner for a project of this scale, with complex requirements and security risks.”
Additional Technical Details
Predictive Module:
This module is fundamental for structuring transaction data from various channels. When transaction data is received from a specific channel, it is transformed into a time series. As a time series, it is processed through an analytical model that generates a table with corresponding predictions for each type of channel transaction, providing a strong foundation for forecasting processes.
This module was built using a combination of Spark and Python. Spark is essential for converting a table with large volumes of information into a time series table, which is then used by Python for consumption and analysis.
If you’d like to learn more about how we can help your company implement solutions to enhance your processes and boost your results, contact us or schedule a meeting with us today!