Power BI and AI integration for efficient sales anomaly detection
Drive business efficiency with data-driven solutions
The client is an IT services provider for a top global beverage manufacturer, with over 12 bottlers sending their sales orders from different sources.
The company was keen on building a data pipeline to capture, store and gain intelligent data and predict anomalies in the sales data being sent into the backend system using machine learning capabilities.
The company’s merchandising systems lacked the infrastructure and intelligence to handle large volumes of data. As a result, the SAP ECC bursts completely exhausted the systems, hampering the functioning of backend data systems.
The client needed to create a pipeline that would allow them to process and store the data, especially for this specific project. The need was immediate, as it pertained to the ongoing project, which was bogged down by diverse issues related to their existing systems.
CoreFlex’s solution envisaged a robust system that could :
The data was stored in SQL databases, which had to be transformed into JSON using a custom tool. The transformed data were then loaded into Power BI, which could be visualized across multiple dimensions and analyzed using different KPIs. The final product was a predictive model that accurately predicted future sales volume anomalies without manual intervention.
A Machine Learning engine was developed to generate trained models based on time series data for sales orders based on different order types, bottlers, and locations.
CoreFlex’s solution was to develop a custom application that the client’s staff would use to record information they were collecting from their system. The system would also be used by them to enter new data as it was generated by their forecasting model.
Once completed, CoreFlex would build a live dashboard that would allow the client to monitor all the activity and send alerts whenever necessary.
The team began by verifying if the business case had been created correctly and that there were no missing data points. Once the validation of the business case was done, we moved on to validate our assumptions. For example, we confirmed our assumptions about how the client would use our product were correct. However, based on research, use-case study, and historic data performance on traditional systems, we found some assumptions were incorrect, so we adjusted them accordingly.
We focused on building a team that could implement these changes successfully. We created a cross-functional team of experts (e.g., design, user experience, and development) so that one person wouldn’t have to take responsibility for multiple areas at once. Another major advantage of using AI ML systems is that it allows them to automate tasks such as data processing that were previously done manually. As a result, using our integrated solution, the client could save money while improving productivity levels.
70%
reduced system downtimes with predictive insights
Enabled sales team with data-driven business decisions on predicting sales volumes
Access to sales data volume and deep insights into sales volume movements and anomalies
Ability to take preventive actions by further streamlining operations at a remote level
Improved partner performance and corresponding support requirements
Improved sales productivity and increase in sales