Popular searches
Not sure where to start?Talk to our team about running sovereign AI in your environment.
↑ ↓ Navigate↵ OpenEsc Close
Vision Banco cut predictive model build time from six months to under a week with H2O Driverless AI, and doubled the number of customers taking up its offers.
Vision Banco's data scientists ran business intelligence the traditional way, with dimensional modeling and ETL into a warehouse. The bank needed to assess credit risk faster and more accurately and move into predictive analytics, but its early attempts, an SPSS model that took a year to build and open source PMML deployment, were still slow and effort-heavy.
The team needed to expand its services, assess credit risk faster and more accurately, and move into predictive analytics such as default and churn prediction. An external consultant first built a model on IBM SPSS, a process that took a year. The team then moved to open source R, H2O, and Openscoring.io to deploy PMML models, but predictive analytics still took heavy time and effort.
Vision Banco adopted H2O Driverless AI to automate the work. Automatic feature engineering, model validation, tuning, selection, and deployment, along with machine learning interpretability, time series, and NLP, cut work that used to take months down to hours or minutes.
The bank ran H2O on an IBM Power System AC922 and changed how it shipped models, moving from PMML to H2O's MOJO scoring pipelines for faster, production-ready deployment.
Model build time dropped from six months or more to under a week. More accurate credit risk models let the bank target offers better, which it estimates is worth millions in additional revenue.
The number of customers buying a credit product doubled, and with sharper targeting the number of customers accepting offers doubled compared with traditional modeling.
“With Driverless AI we save a lot of time, because we accelerate building predictive models with better accuracy. It used to take on average six months or more just to build the models. Now it is done in less than a week.”Ruben DiazData Scientist, Vision Banco
Vision Banco, established in 1992 and headquartered in Asuncion, is the fifth-largest bank in Paraguay and the leader in the country's microfinance segment. It provides specialized financial services to micro, small, and medium enterprises, with a focus on the social and economic development of middle and low income Paraguayans.
The bank offers credit cards, remittances, tax collection, pension contributions, and wire transfers, and holds the largest financial distribution network in the country. More at https://www.visionbanco.com.