Case Study: Akbank achieves a 45% reduction in non-performing loans with Zest AI

Upcoming financial recession could leave a large bank vulnerable

Akbank, one of Turkey’s largest banks, faced economic uncertainty during a period of digital transformation and sought ways to limit losses while sustaining growth. To improve credit decisioning and market responsiveness, Akbank partnered with Zest AI to implement AI-powered underwriting and real-time monitoring.

Zest AI replaced dozens of linear regression models across five product lines using 26x more data, streamlined model development (cutting retrain/deploy time from 10 months to 3 months), and provided early warning of a looming recession six months before the official announcement. As a result, Akbank reduced non-performing loans by 45%, doubled approval rates for applicants with no credit history, and significantly improved underwriting accuracy and agility.

Akbank

Hakan Yilmaz

Head Of Retail Credit And Analytics