KPN Client acceptance platform

Optimizing Customer Acceptance with KYC

By conducting thorough evaluations of the financial situation and history of customers, businesses can better predict whether a customer is a good payer and mitigate the risk of financial losses. This is crucial for companies like KPN, which offer subscriptions and maintain ongoing customer relationships. The application not only helps reduce financial risks but also ensures compliance with consumer credit and debt-related regulations.

KPN processes a significant volume of online and offline transactions daily. With the customer acceptance platform, KPN can make these decisions more efficiently and with greater confidence, which can help improve the overall financial health and stability of the company. The acceptance process takes place in the backend and lasts only a few seconds:

1. A consumer applies for a mobile subscription in a local KPN store.
2. The store employee sends a request to the application with consumer details & subscription request
3. The application performs various checks and sends requests to underlying services based on the provided input.
4. The software sends back a positive or negative finding, including the reason for acceptance or rejection."

This streamlined process not only enhances the customer experience but also helps KPN make informed decisions while adhering to regulatory requirements.

For the device loan check that takes place in the application process, various data sources need to be accessed. When a subscription with a mobile device is requested, variables such as family situation, net income, and housing costs are necessary to know. Based on the input, up to 7 internal and external services are invoked, such as BKR, Preventel, and Experian. To make the calls as fast and efficient as possible, we continuously optimize by grouping long calls, running them asynchronously, and caching where possible.

The backend application is implemented on Amazon Web Services (AWS) and runs in a Kubernetes cluster. Thanks to this architecture, the service has an availability of over 99.999%. The application can scale up and down using a precisely configured Kubernetes Event Driven Autoscaler, which helps maintain availability even during peak loads, such as the launch of the latest iPhone models in 2022.

The application consists of a Django REST API, Kafka-based asynchronous workers for message processing, a Postgres database, and a Redis cluster for caching.

Some key features of the application include:

1. Extreme configurability for business users, allowing them to change the application's behavior in real-time via an admin interface without impacting performance. This helps KPN keep the customer acceptance policy of the application up to date and respond quickly to fraud attempts.

2. A reliable error handling system that collects as much customer information as possible during an application, even when external services experience availability issues.

3. A dashboard for all usage, external API calls, database queries, and business insights such as customer acceptance rates and configurable timeframes.

The application is 4 times faster than the old system due to concurrent calls to external services. This significantly improves the responsiveness to customer input.