A smart use of analytics is helping the bank improve recovery rates, reduce credit losses and optimise customer engagement
The pressure on financial institutions to enhance risk management, reduce non-performing loan ratios and manage customer lifecycles more effectively has never been higher. For Thailand’s Kasikornbank (KBank), this challenge has translated into an opportunity to leverage advanced analytics to transform its collections strategy from a largely reactive system into a proactive, precision-led model.
Precision over volume
At the heart of the bank’s transformation lies its use of internal behavioural data across platforms such as KPLUS to identify the best phone number, time and location to contact each customer. Rather than adopting a brute-force volume approach, KBank engineered a ‘confidence score’ to prioritise contact methods based on usage recency and frequency, thereby increasing successful contact rates while minimising customer dissatisfaction.
Segmentation through smart modelling
The bank’s adoption of a dual-axis framework – willingness to pay (WTP) and ability to pay (ATP) – has allowed it to categorise customers into tailored personas. These personas guide treatment strategies across different product groups, from home loans to unsecured credit. Models have been trained using delinquency history, mobile usage, payment behaviour and income data to improve segmentation accuracy and identify the optimal engagement strategy.
From dashboards to decisions
A full suite of dashboards spanning performance, operations and model monitoring has given KBank real-time insights into customer responsiveness, agent effectiveness and predictive accuracy. This has enabled weekly adjustments and immediate corrective actions where risk thresholds were exceeded or contact protocols deviated from.
The results speak to both strategic vision and execution: the contact rate improved from 68% to 75% within a year and the overall roll rate dropped across multiple lending products. These gains helped reduce expected credit losses (ECL) by THB2.08bn ($60.9bn), significantly surpassing the initial target of THB1.2bn.
For its forward-looking deployment of analytics, KBank was recognised as the winner of the Best Application of Data Analytics title at the Retail Banker International Asia Trailblazer Awards 2025. It also earned a Highly Commended award for Trailblazing Use of AI or Machine Learning in Financial Services, highlighting the technological underpinnings of its strategic evolution.