In the Digital Age, data-based decisions are becoming increasingly important for business. As a result, the demands on the financial sector are also changing. For controlling, this means using predictive analytics to produce more forward-looking analyses and increasingly decision-relevant forecasts instead of focusing on past tense reports.
A further goal is also the extensive automation of routine tasks in order to gain more freedom for strategic analyses and data interpretation. This can be achieved with new tools from the fields of AI and machine learning. And with cloud solutions, they are readily available today with nearly unlimited, on demand computing power.
The prerequisite for forecasts derived with AI support (and other findings) is a clean, comprehensive database. Data management and data integration as core processes of the classic BI environment will therefore remain crucial to success for the next generation of data analytics.
Pilot projects conducted by Jedox have shown that AI-supported analyses within the framework of the Jedox platform very quickly generate business value. More than 50 predefined connectors and the powerful ETL functionality of the Jedox BI platform form the foundation for integrating any data source and large amounts of data.
Automated visualizations allow for easy Master Data Management and to quickly get an overview of data quality during integration. Pilot customers Mitsui Chemicals Europe and Service Master in the USA have subsequently built up extensive, consistent data sets for analysis with the Jedox AI engine.
Automated sales forecast at Mitsui
Mitsui Chemicals Europe upgraded its existing Jedox planning environment with the Jedox Predictive Analytics and AI solution. The project focused specifically on sales forecasting. In order to achieve an even higher forecast accuracy, the project team implemented Jedox’s standard AI module “Predictive Forecast.”
The model, based on the JedoxAIssisted Planning solution, supports Mitsui’s twelve-month rolling sales forecast with AI-generated sales forecasts. After only a few months of operation, the accuracy of the forecast in relation to the actual values increased to up to 95 percent for several of their product groups. Mitsui’s management in the financial area recognized this key advantage in the achieved transparency and granularity of the forecast in addition to the improved efficiency.
The main challenge in the project was the database required for AI analysis. Jedox recommends a database with at least three previous years of data as a basis. However, the planning environment initially only contained data from 1.5 years. In order to increase the data set to three years, the team therefore adjusted further legacy data in the SAP upstream system using mapping with the help of the Jedox Integrator and migrated it to Jedox.