Return to page

İşbank selects Driverless AI platform from H2O.ai for Income Prediction, Cash Forecasting & Check Default Prediction applications

June 22, 2021

H2O.ai, a leading AI technology software company, has announced that İşbank, officially Türkiye İş Bankası, the largest private bank in Turkey, has selected H2O’s Driverless AI platform to upgrade its Income Prediction, Cash Forecasting and Check Default Prediction applications, to improve the accuracy of the bank’s business planning with on-going AI projects.

In close partnership with İşbank’s San Francisco-based innovation arm Maxitech, who is leading the initiative, İşbank has agreed to adopt the H2O Driverless AI platform after a successful project which saw: machine learning (ML) performance improvements, revenue increases and cost reductions on existing application development processes.

Ümit Bilik, Chief Data Officer, Head of Data Management & Artificial Intelligence at İşbank said: “We are delighted to have selected H2O.ai through the work with Maxitech for this important step towards empowering our AI capabilities. Our selection decision was driven by the proven creation of successful high-performance machine learning models, demonstrable efficiency improvements and H2O.ai’s technical and domain expertise.”

Andrew Morris, Senior Vice President, EMEA, H2O.ai said, “The decision by İşbank to use Driverless AI is another shining example of how a progressive-thinking finance organisation in EMEA is embracing innovative AI and ML technologies to improve the accuracy of its business planning predictions, and be even more competitive in an increasingly challenging market. We look forward to working closely with İşbank, and its innovation arm Maxitech, on the next stage of this IT business transformational initiative.”

During the initial project stage, İşbank, saw the following results: ·

Income Prediction Results:

  • 8% performance increase with Automatic Machine Learning (AutoML).

Cash Forecasting Results:

Increased revenue by optimized cash management across 4000 piloted ATMs with up to 50% reduction in cash forecast deviation

Reduced re-training effort of models from five days to six hours

With AutoML, a decrease in the time to develop new models from 2 months to 2.5 hours

Check Default Prediction Results:

  • With AutoML, a decrease in the time to develop new models from 2 months to 2.5 hours
  • Teams with in-depth AI knowledge can learn/use model in shorter timeframes

In addition to the primary selection agreement by İşbank, Maxitech introduced İşbank’s software development subsidiary Softtech which is one of the leading tech companies in Turkey to H2O.ai. Both parties worked together on assessing how H2O.ai’s Driverless AI platform would enable the business unit to digitally transform its customers in the Turkish marketplace into more AI driven organisations. In 2021, Softtech strategically positions as the visionary for Automated Machine Learning Platform needs in the Turkish market.